CN112818671A

CN112818671A - Text information processing method and device, storage medium and processor

Info

Publication number: CN112818671A
Application number: CN201911120523.9A
Authority: CN
Inventors: 周鑫; 张雅婷; 孙常龙; 张琼; 司罗
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2019-11-15
Filing date: 2019-11-15
Publication date: 2021-05-18

Abstract

The invention discloses a text information processing method, a text information processing device, a storage medium and a processor. Wherein, the method comprises the following steps: acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case. The method solves the technical problem of low efficiency of processing the case in the litigation scene because the litigation result of the case and the legal information used cannot be predicted in the litigation scene.

Description

Text information processing method and device, storage medium and processor

Technical Field

The present invention relates to the field of text information processing, and in particular, to a method and an apparatus for processing text information, a storage medium, and a processor.

Background

At present, with the enhancement of legal consciousness of people, cases are more and more, and particularly, the number of civil cases is high, and the highest court encourages more civil cases to enter a multi-mediation dispute stage. However, currently, manual mediation is usually performed according to actual situations of cases, time consumption is long, and a scheme for predicting litigation results of cases and legal information used in litigation scenes is not provided, so that the technical problem that efficiency of managing cases is low exists.

In order to solve the problem that the case processing efficiency in the litigation scene is low because the litigation result of the case and the legal information used in the case cannot be predicted in the litigation scene, an effective solution is not provided at present.

Disclosure of Invention

The embodiment of the invention provides a method, a device, a storage medium and a processor for processing text information, which are used for at least solving the technical problem of low efficiency of processing cases in litigation scenes due to the fact that the litigation results of the cases and the legal information used in the cases cannot be predicted in the litigation scenes.

According to an aspect of the embodiments of the present invention, a method for processing text information is provided. The method comprises the following steps: acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.

According to another aspect of the embodiment of the invention, a method for processing text information is also provided. The method comprises the following steps: displaying target text information on a text display interface, wherein the target text information is used for describing case information of cases in litigation scenes; and displaying legal information and litigation prediction information of the case on a text display interface, wherein the legal information and the litigation prediction information of the case are results of prediction processing on target text information based on a multi-task prediction model, and the multi-task prediction model is obtained by training text information of case information for describing case samples in litigation scenes and legal information samples of the case samples.

According to another aspect of the embodiment of the invention, a method for processing case information is also provided. The method comprises the following steps: acquiring case information of a case, wherein the case information of the case comprises at least one of the following: appeal shape text, answer shape text, law and regulation text, evidence text and judgment text; predicting case information of a case based on a multitask prediction model to obtain legal information and litigation prediction information of the case, wherein the multitask prediction model is a neural network model and is used for coding different types of files in the case information of the case; and outputting legal information and litigation prediction information of the case.

According to another aspect of the embodiment of the invention, a text information processing device is also provided. The device includes: the first acquisition unit is used for acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; the first processing unit is used for predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and the first output unit is used for outputting legal information and litigation prediction information of the case.

According to another aspect of the embodiment of the invention, a text information processing device is also provided. The device includes: the first display unit is used for displaying target text information on a text display interface, wherein the target text information is used for describing case information of cases in litigation scenes; and the second display unit is used for displaying the legal information and the litigation prediction information of the case on the text display interface, wherein the legal information and the litigation prediction information of the case are the result of prediction processing on the target text information based on the multi-task prediction model, and the multi-task prediction model is obtained by training the text information of the case information for describing the case sample in the litigation scene and the legal information sample of the case sample.

According to another aspect of the embodiment of the invention, a device for processing case information is also provided. The device includes: a second obtaining unit, configured to obtain case information of a case, where the case information of the case includes at least one of: appeal shape text, answer shape text, law and regulation text, evidence text and judgment text; the second processing unit is used for predicting the case information of the case based on the multitask prediction model to obtain legal information and litigation prediction information of the case through prediction, wherein the multitask prediction model is a neural network model and is used for coding different types of files in the case information of the case; and the second output unit is used for outputting the legal information and litigation prediction information of the case.

According to another aspect of the embodiments of the present invention, there is also provided a storage medium. The storage medium includes a stored program, wherein the program, when executed, controls an apparatus in which the storage medium is located to perform the steps of: acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.

According to another aspect of the embodiments of the present invention, there is also provided a processor. The processor is used for running the program, wherein the program executes the following steps: acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.

According to another aspect of the embodiment of the invention, the invention also provides the mobile terminal. The mobile terminal includes: a processor; the transmission device is connected with the processor and is used for acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; and a memory coupled to the processor for providing instructions to the processor for processing the following processing steps: predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.

In the embodiment of the invention, target text information is obtained, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case. That is, the case prediction information and the legal information are jointly predicted through the multi-task prediction model trained in advance, so that the purpose of assisting the case examination is achieved, the efficiency of processing the case in the litigation scene is improved, and the technical problem that the efficiency of processing the case in the litigation scene is low because the litigation result of the case and the used legal information cannot be predicted in the litigation scene is solved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

fig. 1 shows a hardware configuration block diagram of a computer terminal (or mobile device) for implementing a processing method of text information;

FIG. 2 is a flow chart of a method of processing text information according to an embodiment of the invention;

FIG. 3 is a flow chart of another method of processing text information according to an embodiment of the invention;

FIG. 3a is a schematic diagram of an interaction in another method for processing text information according to an embodiment of the present invention;

FIG. 4 is a flow chart of a method for processing case information according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of a system for processing text information in accordance with an embodiment of the present invention;

FIG. 6 is a schematic diagram of the architecture of a multitasking model according to an embodiment of the present invention;

FIG. 7 is a schematic diagram of a network architecture of a multitasking model according to an embodiment of the present invention;

FIG. 8 is a schematic diagram of an apparatus for processing text information according to an embodiment of the present invention;

FIG. 9 is a schematic view of another text information processing apparatus according to an embodiment of the present invention;

FIG. 10 is a schematic diagram of a case information processing apparatus according to an embodiment of the present invention; and

fig. 11 is a block diagram of a mobile terminal according to an embodiment of the present invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

First, some terms or terms appearing in the description of the embodiments of the present application are applicable to the following explanations:

a Term Frequency-Inverse text Frequency index (Term Frequency-Inverse text Frequency, abbreviated as TF-IDF) is a common weighting technique used for information retrieval and data mining;

a Continuous Bag-of-Words (CBOW) model, wherein the training input of the CBOW model is a word vector corresponding to a word related to the context of a certain characteristic word, and the output is the word vector of the specific word;

a Skip-gram neural network model, wherein the input of the Skip-gram neural network model is a word vector of a specific word, and the output of the Skip-gram neural network model is a context word vector corresponding to the specific word;

a novel deep contextualized word representation algorithm (ELMo) for modeling complex features (such as syntax and semantics) of words and changes of words in Language context (i.e., modeling ambiguous words);

the Self-attention Mechanism (Self-attention Mechanism) is applied to various tasks of natural language processing based on deep learning.

Example 1

There is also provided, in accordance with an embodiment of the present invention, a method embodiment of a method for processing textual information, it being noted that the steps illustrated in the flowchart of the figure may be performed in a computer system such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flowchart, in some cases the steps illustrated or described may be performed in an order different than here.

The method provided by the first embodiment of the present application may be executed in a mobile terminal, a computer terminal, or a similar computing device. Fig. 1 shows a hardware configuration block diagram of a computer terminal (or mobile device) for implementing a text information processing method. As shown in fig. 1, the computer terminal 10 (or mobile device 10) may include one or more (shown as 102a, 102b, … …, 102 n) processors 112 (the processors 112 may include, but are not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA), memory 114 for storing data, and a transmission device for communication functions. Besides, the method can also comprise the following steps: a display, an input/output interface (I/O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of the I/O interface), a network interface, a power source, and/or a camera. It will be understood by those skilled in the art that the structure shown in fig. 1 is only an illustration and is not intended to limit the structure of the electronic device. For example, the computer terminal 10 may also include more or fewer components than shown in FIG. 1, or have a different configuration than shown in FIG. 1.

It should be noted that the one or more processors 112 and/or other data processing circuitry described above may be referred to generally herein as "data processing circuitry". The data processing circuitry may be embodied in whole or in part in software, hardware, firmware, or any combination thereof. Further, the data processing circuit may be a single stand-alone processing module, or incorporated in whole or in part into any of the other elements in the computer terminal 10 (or mobile device). As referred to in the embodiments of the application, the data processing circuit acts as a processor control (e.g. selection of a variable resistance termination path connected to the interface).

The memory 114 may be used to store software programs and modules of application software, such as program instructions/data storage devices corresponding to the processing method of text information in the embodiment of the present invention, and the processor 112 executes various functional applications and data processing by executing the software programs and modules stored in the memory 114, that is, implementing the processing method of text information of the application program. The memory 114 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 114 may further include memory located remotely from the processor 112, which may be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmission device is used for receiving or transmitting data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device includes a Network adapter (NIC) that can be connected to other Network devices through a base station to communicate with the internet. In one example, the transmission device may be a Radio Frequency (RF) module, which is used for communicating with the internet in a wireless manner.

The display may be, for example, a touch screen type Liquid Crystal Display (LCD) that may enable a user to interact with a user interface of the computer terminal 10 (or mobile device).

It should be noted here that in some alternative embodiments, the computer device (or mobile device) shown in fig. 1 described above may include hardware elements (including circuitry), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware and software elements. It should be noted that fig. 1 is only one example of a particular specific example and is intended to illustrate the types of components that may be present in the computer device (or mobile device) described above.

In the operating environment shown in fig. 1, the present application provides a method for processing text information as shown in fig. 2. It should be noted that the text information processing method of this embodiment may be executed by the mobile terminal of the embodiment shown in fig. 1.

Fig. 2 is a flowchart of a text information processing method according to an embodiment of the present invention. As shown in fig. 2, the method comprises the steps of:

step S202, target text information is obtained, wherein the target text information is used for describing case information of cases in litigation scenes.

In the technical solution provided by step S202 of the present invention, the case in the litigation scene may be a case that needs to be disputed, for example, a civil case. The case information of this embodiment may include textual materials such as complaint form text, answer form text, legal and legal provisions text, evidence text, and decision text, which are required in the course of reviewing the case. The target text information of the embodiment is used for describing case information of cases in the litigation scene, and may include the case information.

And step S204, performing prediction processing on the target text information based on the multi-task prediction model, and predicting to obtain legal information and litigation prediction information of the case.

In the technical solution provided in step S204 of the present invention, after the target text information is obtained, the target text information is predicted based on a multitask prediction model, and legal information and litigation prediction information of a case are obtained by prediction, where the multitask prediction model is trained through text information of case information describing case samples in litigation scenes and legal information samples of the case samples, and a model for predicting legal information and litigation prediction information of different cases is obtained.

In this embodiment, after the target text information is obtained, the target text information may be predicted through a pre-trained multi-task prediction model, which may be a neural network model, and may be obtained by training case samples collected in a litigation scene, for example, by pre-collecting text information used to describe the case information of the case sample, the litigation prediction information of the case sample, and legal information samples of the case samples associated with the litigation prediction information of the case sample in the litigation scene, selecting an algorithm strategy for multi-task learning, and training the sub-neural network model to obtain the multi-task prediction model, where the sub-neural network model may be an initially established detection model. The legal information of the case obtained by the prediction processing through the task prediction model of the embodiment is highly related to the litigation prediction information, wherein the legal information can be legal regulations, for example, legal terms, and the litigation prediction information of the case, that is, the result of the complaint, and the legal information of the case is highly related to the litigation prediction information.

Optionally, when the sub-neural network model is trained through the text information of the case sample, the litigation prediction information of the case sample and the law information sample, the text information of the case sample, the litigation prediction information of the case sample and the legal information of the case sample associated with the litigation prediction information of the case sample can be preprocessed according to the algorithms of distribution consistency algorithm, denoising, sampling and the like, then the features used for training the sub-neural network model are obtained through feature extraction, feature transformation, feature normalization, feature combination and the like from the preprocessed data, and further the features are processed through the optimization algorithm, the hypothesis function, the loss function, the decision boundary, the convergence speed, the iteration strategy and the like to obtain the label of the relationship between the text information of the case and the legal information and the litigation prediction information of the case, and determining parameters of the sub-neural network model through the label, and determining the multitask prediction model through the parameters. The embodiment can also perform cross validation, target evaluation, overfitting, underfitting and other evaluations on the multi-task prediction model obtained by training finally, so that a model for predicting legal information and litigation prediction information of different cases is obtained.

Step S206, legal information and litigation prediction information of the case are output.

In the technical solution provided in step S206 of the present invention, after the target text information is predicted based on the multitask prediction model, and the legal information and litigation prediction information of the case are obtained through prediction, the legal information and litigation prediction information of the case are output, and the legal information and litigation prediction information can be output to a display interface, so as to serve as a reference result for case adjustment, and thus, a case moderator can make a case mediation strategy according to the predicted legal information and litigation prediction information. The embodiment can also be flexibly applied to other judicial auxiliary scenes, for example, legal information and litigation prediction information obtained through prediction are provided for lawyers and parties to be referred, so that the examination of cases is assisted, and the processing efficiency of the cases is improved.

Through the steps S202 to S206, target text information is obtained, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case. That is, the case prediction information and the legal information are jointly predicted through the multi-task prediction model trained in advance, so that the purpose of assisting the case examination is achieved, the efficiency of processing the case in the litigation scene is improved, and the technical problem that the efficiency of processing the case in the litigation scene is low because the litigation result of the case and the used legal information cannot be predicted in the litigation scene is solved.

As an optional implementation manner, before processing the target text information based on the multitask prediction model to obtain the legal information and litigation prediction information of the case in step S204, the method further includes: training through the text information of the legal information sample and the case sample to obtain a first task prediction model, wherein the multi-task prediction model comprises a first task prediction model, and the first task prediction model is used for predicting the legal information of the case; and training the legal information samples, the text information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the multi-task prediction model comprises the second task prediction model, and the second task prediction model is used for predicting litigation prediction information of the case.

In this embodiment, before the target text information is processed based on the multi-task prediction model to obtain legal information and litigation prediction information of a case, the multi-task prediction model needs to be trained. Because the legal information and the litigation prediction information of the case are highly correlated, the embodiment can adopt an algorithm strategy of multi-task learning, train a first task prediction model for predicting the legal information of the case and a second task prediction model for predicting the litigation prediction information of the case, and the first task prediction model and the second task prediction model of the embodiment are also correlated, so that the characteristics can be shared, and the accuracy of the results of the legal information and the litigation prediction information of the case can be improved.

Optionally, in the embodiment, the legal information sample of the case sample and the text information of the case sample are trained to obtain the first mission prediction model, wherein the legal information sample may be a formal text of legal provision, the text information of the case sample may be the case information of the case sample, and may be text materials such as appeal shapes, answer shapes, decision books and the like of different cases, and the more the text materials are, the better the first mission prediction model is; in this embodiment, the legal information output by the first task prediction model may be used as input information of the second task prediction model during training, so that the second task prediction model is obtained by training the legal information sample, the text information of the case sample, and the legal information output by the first task prediction model, thereby ensuring that the trained second task prediction model is highly related to the first task prediction model.

As an optional implementation, the method further comprises: coding the legal information sample to obtain a first coding result; coding the text information of the case sample to obtain a second coding result; coding discrete characteristics of the text information of the case sample to obtain a third coding result; training through the text information of the legal information sample and the case sample to obtain a first task prediction model, wherein the first task prediction model comprises the following steps: training through the first coding result, the second coding result and the third coding result to obtain a first task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the first task prediction model; training through the legal information samples, the text information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through the first coding result, the second coding result, the third coding result and legal information output by the first task prediction model to obtain a second task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the second task prediction model.

In this embodiment, the legal information sample may be encoded to obtain a first encoding result, where the first encoding result is a result in the input layer, and the legal information sample may be encoded in a legal information encoding layer of the input layer to obtain the first encoding result; the embodiment can also encode the text information of the case sample to obtain a second encoding result, wherein the second encoding result is also the result of the input layer, and the text information of the case sample can be encoded in the case text information encoding layer of the input layer to obtain a second encoding result; the embodiment can also acquire the discrete features of the text information of the case sample, encode the discrete features to obtain a third encoding result, the third encoding result is also the result of the input layer, and encode the discrete features of the text information in the discrete feature input layer of the input layer to obtain a third encoding result.

Optionally, in this embodiment, when the first task prediction model is obtained by training the text information of the legal information sample and the case sample, the first coding result, the second coding result, and the third coding result are used as the results of the input layer of the first task prediction model, and the first task prediction model is obtained by training.

Optionally, when the second task prediction model is obtained by training the legal information samples, the text information of the case samples, and the legal information output by the first task prediction model, the first encoding result, the second encoding result, and the third encoding result may be the result of the input layer of the second task prediction model, and the second task prediction model is obtained by training.

As an optional implementation, encoding the legal information sample to obtain the first encoding result includes: performing word segmentation on a legal information sample to obtain a first word vector sequence; the first word vector sequence is determined as a first encoding result.

In this embodiment, the legal information sample may be encoded in the legal information encoding layer to obtain a first encoding result, and the legal information sample may be participled to obtain a plurality of participles. The word vector of the embodiment is pre-trained by using the ELMo model, and compared with the prior fixed word vector, the ELMo model considers context information, and different context word vectors are different, so that the problems of ambiguous words and the like can be solved. In this embodiment, after the word vector is trained through the corpus after word segmentation, all legal terms in the legal information sample can be encoded into the first word vector sequence, and then the first word vector sequence is determined as the first encoding result.

As an optional implementation manner, the encoding the text information of the case sample, and obtaining the second encoding result includes: segmenting words of the text information of the case sample to obtain a second word vector sequence; and coding the second word vector sequence to obtain a second coding result.

In this embodiment, the text information of the case sample may be encoded in the case text information encoding layer to obtain a second encoding result, the text information of the case sample may be participled to convert the text information into a second word vector sequence, and then the second word vector sequence is encoded, and the second word vector may be encoded by using a cudnn/gru model and a Self-attribute mechanism to obtain a second encoding result, which is also the case text encoding finally.

As an optional implementation manner, the encoding discrete features of the text information of the case sample, and obtaining the third encoding result includes: and coding the discrete features through a multilayer perceptron to obtain a third coding result with continuous semantics.

According to the embodiment, the discrete features of the text information of the case sample can be coded in the discrete feature input layer, the multilayer perceptron can be adopted for code conversion of the discrete features, the discrete feature vectors can be converted into the continuous semantic space, and a third coding result is obtained, so that the method is more suitable for training of the deep learning model.

As an optional implementation manner, the training by the first encoding result, the second encoding result, and the third encoding result to obtain the first task prediction model includes: performing fusion coding on the second coding result and the third coding result to obtain a fourth coding result; splicing the fourth coding result and the first coding result to obtain a fifth coding result; and training through a fifth coding result to obtain the first task prediction model, wherein the fifth coding result is a result of an intermediate layer of the first task prediction model.

In this embodiment, when the first task prediction model is obtained by training the first encoding result, the second encoding result, and the third encoding result, the second encoding result and the third encoding result may be fusion encoded in the french encoding layer, for example, the second encoding result of the case text information encoding layer and the third encoding result of the discrete feature input layer may be fusion encoded by using Self-attention and highway strategies, so as to obtain the third encoding result. After the second coding result and the third coding result are subjected to fusion coding to obtain a fourth coding result, the fourth coding result and the first coding result may be spliced, for example, the fourth coding result and the first coding result of the french information coding layer are spliced to obtain a fifth coding result, the fifth coding result is a result of an intermediate layer of the first task prediction model, and the intermediate layer may be the french coding layer. That is, in this embodiment, the case text information coding layer and the discrete feature input layer are first fused and coded through the Self-attention and highway strategies, and then spliced with the french information coding layer, so as to obtain the french coding layer.

As an optional implementation, the training by the fifth encoding result to obtain the first task prediction model includes: and determining an output layer of the first task prediction model according to the fifth coding result and the objective function to obtain the first task prediction model.

In this embodiment, when the first task prediction model is obtained by training through the fifth encoding result, the output layer of the first task prediction model may be determined through the fifth encoding result and the objective function, so as to obtain the first task prediction model. Optionally, the output layer of the first task prediction model in this embodiment may be a french output layer, the objective function may be a Softmax function, and the french information of the french output layer is obtained through prediction by the fifth encoding result of the french encoding layer and the Softmax function.

As an optional implementation manner, the training by the first encoding result, the second encoding result, the third encoding result, and the legal information output by the first task prediction model, and the obtaining of the second task prediction model includes: classifying the first coding result through a pre-constructed classifier of the legal information sample to obtain a classification result; splicing the classification result, the intermediate classification result used for obtaining the classification result, the second coding result and the third coding result to obtain a splicing result; performing fusion coding on the splicing result to obtain a sixth coding result, wherein the sixth coding result is the result of the middle layer of the second task prediction model; and training through the sixth coding result and legal information output by the first task prediction model to obtain a second task prediction model.

In this embodiment, when the second task prediction model is obtained by training the first encoding result, the second encoding result, the third encoding result, and the legal information output by the first task prediction model, a classifier of a legal information sample may be constructed first, where the classifier may correspond to a legal label encoding layer, may be an attribute label classifier, may be a multi-label classifier, and may be constructed manually. According to the embodiment, the first coding result can be classified through the constructed classifier to obtain the classification result, and the classification result, the intermediate classification result used for obtaining the classification result, the second coding result and the third coding result are spliced to obtain the splicing result. Optionally, in this embodiment, a law enforcement result encoding layer is determined according to an intermediate result used for obtaining the classification result, and the classification result of the law label encoding layer, the intermediate classification result of the law enforcement result encoding layer, the second encoding result of the case text information encoding layer, and the third encoding result of the discrete feature input layer are spliced, that is, the four items of the law label encoding layer, the law enforcement result encoding layer, the case text information encoding layer, and the discrete feature input layer are spliced to obtain a splicing result, and then the splicing result is fusion encoded, and a highway policy may be used to fusion encode the splicing result to obtain a sixth encoding result, where the sixth encoding result is a result of an intermediate layer of the second task prediction model, and the intermediate layer may be a complaint result encoding layer.

After the splicing result is subjected to fusion coding to obtain a sixth coding result, the sixth coding result and legal information output by the first task prediction model can be trained to obtain a second task prediction model, namely, the prediction legal information can be used as input information of predicted litigation prediction information, and the training of the model is added in the last layer to further obtain the second task prediction model.

As an optional implementation, the training by the sixth encoding result and the legal information output by the first task prediction model to obtain the second task prediction model includes: determining an output layer of the second task prediction model according to the sixth coding result and the objective function; and training an output layer of the second task prediction model through legal information output by the first task prediction model to obtain the second task prediction model.

In this embodiment, when the second task prediction model is obtained by training the sixth encoding result and the legal information output by the first task prediction model, the output layer of the second task prediction model may also be determined by the sixth encoding result and the objective function, so as to obtain the second task prediction model. Alternatively, the output layer of the second task prediction model of this embodiment may be an output layer of litigation prediction information, and the target function may be a Softmax function, and the litigation prediction information is predicted by the sixth encoding result of the appeal result encoding layer and the Softmax function.

As an optional implementation manner, training the output layer of the second task prediction model through legal information output by the first task prediction model, and obtaining the second task prediction model includes: acquiring a first loss function corresponding to legal information output by a first task prediction model; acquiring a second loss function corresponding to an output layer of the second task prediction model; and training an output layer of the second task prediction model through the first loss function and the second loss function to obtain the second task prediction model.

In the embodiment, legal information output by the first task prediction model corresponds to a first loss function, an output layer of the second task prediction model corresponds to a second loss function, and the optimization of the model of the embodiment aims to synthesize the first loss function and the second loss function and train the output layer of the second task prediction model to obtain the second task prediction model, so that the first task prediction model and the second task prediction model are highly related.

As an optional implementation, the method further comprises: extracting the statistical characteristics of words and the vector characteristics of the words from the text information of case samples; training through the text information of the legal information sample and the case sample to obtain a first task prediction model, wherein the first task prediction model comprises the following steps: training through legal information samples, the statistical characteristics of words and the vector characteristics of the words to obtain a first task prediction model; training through the legal information samples, the text information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through legal information samples, the statistical characteristics of the words, the vector characteristics of the words and the legal information output by the first task prediction model to obtain a second task prediction model.

The embodiment can also establish a text preprocessing and pre-training model, and the model is used for preprocessing target text information, so that the training of the multi-task prediction model is facilitated, for example, text materials such as the appealing text, the answer-like text and the evidence text are preprocessed, wherein the appealing text and the answer-like text are the most important case situation analysis materials, and the evidence text submitted in the original way is also an important feature source for case description and anti-answer text. Optionally, the embodiment extracts the statistical features of the words and the vector features of the words from the text information of the case sample, wherein the statistical features may be TF-IDF.

Optionally, the preprocessing step of this embodiment includes: performing word segmentation processing on the text information, wherein any mainstream word segmentation tool can be used and a legal dictionary needs to be added; part-of-speech tagging, i.e., tagging each participle with part-of-speech tags, such as verbs, nouns, conjunctions, helpers, etc.; entity identification, for example, entity identification such as a person name, a place name, and an organization name is performed, which is not limited herein; excavating statistic characteristics of TF-TDF, TF and the like; the word vector pre-training is performed, when the word vector pre-training is performed, the word vectors can be respectively trained according to different cases, and text information such as appeal, answer, judgment and the like of different cases can be used, wherein the more the text information is, the better the text information is. The word segmentation is performed on the text information of different cases, the word vector is trained through the linguistic data after the word segmentation, and the word vector can be trained by adopting a CBOW model, a Skip-gram model, an ELMo model, a bert model and the like without any limitation. Preferably, the embodiment trains the word vectors using the ELMo model to be able to solve the ambiguous word problem and the like.

After the statistical characteristics of words and the vector characteristics of the words are extracted from the text information of case samples, when a first task prediction model is obtained through training of the text information of legal information samples and case samples, the statistical characteristics of the legal information samples and the words and the vector characteristics of the words can be trained to obtain the first task prediction model, when a second task prediction model is obtained through training of the text information of the legal information samples and the case samples and the legal information output by the first task prediction model, the statistical characteristics of the legal information samples and the words and the vector characteristics of the words and the legal information output by the first task prediction model can be trained to obtain the second task prediction model.

Under the operating environment shown in fig. 1, the present application further provides a method for processing text information as shown in fig. 3. It should be noted that the text information processing method of this embodiment may be executed by the mobile terminal of the embodiment shown in fig. 1.

Fig. 3 is a flowchart of another text information processing method according to an embodiment of the present invention. As shown in fig. 3, the method comprises the steps of:

step S302, target text information is displayed on a text display interface, wherein the target text information is used for describing case information of cases in litigation scenes.

In the technical solution provided by step S302 of the present invention, the case in the litigation scene may be a case that needs to be disputed. The case information of this embodiment may include textual materials such as complaint form text, answer form text, legal and legal provisions text, evidence text, and decision text, which are required in the course of reviewing the case. And displaying target text information on a text display interface, wherein the target text information is used for describing case information of cases in the litigation scene, and the case information can be included.

And S304, displaying the legal information and litigation prediction information of the case on a text display interface, wherein the legal information and the litigation prediction information of the case are the results of prediction processing on the target text information based on the multi-task prediction model.

In the technical solution provided in step S304 of the present invention, after the target text information is displayed on the text display interface, the legal information and litigation prediction information of the case are displayed on the text display interface, wherein the multitask prediction model is obtained by training the text information of the case information for describing the case sample in the litigation scene and the legal information sample of the case sample.

In this embodiment, after the target text information is displayed on the text display interface, the target text information may be predicted through a pre-trained multi-task prediction model, which may be a neural network model, and may be obtained by training case samples collected in a litigation scene, for example, by pre-collecting text information used to describe the case information of the case sample, the litigation prediction information of the case sample, and legal information samples of the case sample associated with the litigation prediction information of the case sample in the litigation scene, selecting an algorithm strategy for multi-task learning, and training the sub-neural network model to obtain the multi-task prediction model, where the sub-neural network model may be an initially established detection model. Legal information and litigation prediction information of a case obtained by prediction processing by the task prediction model of this embodiment are highly correlated.

Specifically, as shown in fig. 3a, fig. 3a is a schematic diagram of interaction in another text information processing method according to an embodiment of the present invention, a user may drag text materials such as a prosecution form text, a quiz form text, a legal and legal regulation text, an evidence text, and a decision text, which are required in a case trial process, into an "add" text box, perform litigation prediction on target text information by clicking an "information processing" key, and finally generate legal information and litigation prediction information related to the case, where the litigation prediction information may include litigation success probability, risk information, potential problems, success probability of upper appeal after a complaint, or related information of upper complaints.

After the target text information is subjected to prediction processing based on the multi-task prediction model and legal information and litigation prediction information of a case are obtained through prediction, the legal information and the litigation prediction information of the case are displayed on a text display interface, so that the legal information and the litigation prediction information obtained through prediction are used as reference results for case trial, and a case moderator can make a case mediation strategy according to the predicted legal information and the litigation prediction information. The embodiment can also be flexibly applied to other judicial auxiliary scenes, for example, legal information and litigation prediction information obtained through prediction are provided for lawyers and parties to be referred, so that the examination of cases is assisted, and the processing efficiency of the cases is improved.

Under the operating environment shown in fig. 1, the present application further provides a case information processing method as shown in fig. 4. It should be noted that the case information processing method of this embodiment may be executed by the mobile terminal of the embodiment shown in fig. 1.

Fig. 4 is a flowchart of a case information processing method according to an embodiment of the present invention. As shown in fig. 4, the method comprises the steps of:

step S402, acquiring case information of the case.

In the technical solution provided by step S402 of the present invention, the case information of the case includes at least one of the following: appeal text, answer form text, legal and legal text, evidence text, and decision text.

In this embodiment, case information may include textual material such as complaint form text, answer form text, legal and legal text, evidence text, and decision text, which are required in the course of reviewing a case.

And S404, performing prediction processing on the case information of the case based on a multitask prediction model to obtain legal information and litigation prediction information of the case through prediction, wherein the multitask prediction model is a neural network model and is used for encoding different types of files in the case information of the case.

In the technical solution provided in step S404 of the present invention, after the case information of the case is obtained, the case information of the case is predicted based on a multitask prediction model, and the legal information and litigation prediction information of the case are obtained by prediction, where the multitask prediction model is a neural network model and is used for coding different types of files in the case information of the case.

In this embodiment, after obtaining the case information of the case, the case information of the case may be predicted through a pre-trained multitask prediction model, which may be a neural network model, and different types of files in the case information of the case may be encoded to obtain legal information and litigation prediction information of the case.

Optionally, when the multitask prediction model encodes different types of files in case information of a case, the text in the file may be segmented in a legal information encoding layer of the multitask prediction model to obtain a word vector sequence, the text in the file may be segmented in a case text information encoding layer to be converted into a word vector sequence, and then the word vector sequence may be encoded through cudnn/gru + Self-entry to obtain a case text encoding, and the text in the file may be encoded and converted by using a multilayer sensor network in a discrete feature input layer to obtain legal information and litigation prediction information of the case.

Optionally, the embodiment selects an algorithm strategy for multi-task learning by collecting text information of case information used for describing case samples, litigation prediction information of the case samples and legal information samples of the case samples associated with the litigation prediction information of the case samples in a litigation scene in advance, and trains the sub-neural network model to obtain the multi-task prediction model, where the sub-neural network model may be an initially established detection model. The legal information of the case obtained by the prediction processing through the task prediction model of the embodiment is highly related to the litigation prediction information, wherein the legal information can be legal regulations, for example, legal terms, and the litigation prediction information of the case, that is, the result of the complaint, and the legal information of the case is highly related to the litigation prediction information.

Step S406, legal information and litigation prediction information of the case are output.

In the technical solution provided in step S406 of the present invention, after the case information of the case is predicted based on the multitask prediction model, and the legal information and litigation prediction information of the case are obtained through prediction, the legal information and litigation prediction information of the case are output, and the legal information and litigation prediction information of the case can be output to a display interface to be used as a reference result for case trial.

As an optional implementation manner, in step S404, before the case information is subjected to prediction processing based on the multitask prediction model, and legal information and litigation prediction information of the case are predicted, the method further includes: training through legal information samples of case samples and case information of the case samples to obtain a first task prediction model, wherein the multi-task prediction model comprises a first task prediction model, and the first task prediction model is used for predicting the legal information of the cases; and training the legal information samples, the case information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the multi-task prediction model comprises the second task prediction model, and the second task prediction model is used for predicting litigation prediction information of the case.

In this embodiment, the multi-tasking prediction model needs to be trained before the case information is processed based on the multi-tasking prediction model to predict legal information and litigation prediction information of the case. Because the legal information and the litigation prediction information of the case are highly correlated, the embodiment can adopt an algorithm strategy of multi-task learning, train a first task prediction model for predicting the legal information of the case and a second task prediction model for predicting the litigation prediction information of the case, and the first task prediction model and the second task prediction model of the embodiment are also correlated, so that the characteristics can be shared, and the accuracy of the results of the legal information and the litigation prediction information of the case can be improved.

Optionally, the embodiment trains through the legal information sample of the case sample and the case information of the case sample to obtain a first task prediction model; in this embodiment, the legal information output by the first task prediction model may be used as input information of the second task prediction model during training, so that the second task prediction model is obtained by training the legal information sample, the case information of the case sample, and the legal information output by the first task prediction model, thereby ensuring that the trained second task prediction model is highly related to the first task prediction model.

As an optional implementation, the method further comprises: coding the legal information sample to obtain a first coding result; coding case information of the case sample to obtain a second coding result; coding discrete characteristics of case information of the case sample to obtain a third coding result; training through legal information samples and case information of the case samples to obtain a first task prediction model, wherein the first task prediction model comprises the following steps: training through the first coding result, the second coding result and the third coding result to obtain a first task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the first task prediction model; training through the legal information sample, the case information of the case sample and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through the first coding result, the second coding result, the third coding result and legal information output by the first task prediction model to obtain a second task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the second task prediction model.

In this embodiment, the legal information sample may be encoded to obtain a first encoding result, where the first encoding result is a result in the input layer, and the legal information sample may be encoded in a legal information encoding layer of the input layer to obtain the first encoding result; the embodiment can also encode the case information of the case sample to obtain a second encoding result, the second encoding result is also the result of the input layer, and the case information of the case sample can be encoded in the case text information encoding layer of the input layer to obtain a second encoding result; the embodiment can also acquire the discrete features of the text information of the case sample, encode the discrete features to obtain a third encoding result, the third encoding result is also the result of the input layer, and encode the discrete features of the case information in the discrete feature input layer of the input layer to obtain a third encoding result.

Optionally, in this embodiment, when the first task prediction model is obtained by training the case information of the legal information sample and the case sample, the first coding result, the second coding result, and the third coding result are used as the results of the input layer of the first task prediction model, and the first task prediction model is obtained by training.

According to the case processing method and the case processing system, the case prediction information and the legal information are jointly predicted for the submitted cases through the multi-task prediction model trained in advance, so that the purpose of assisting in case examination is achieved, the efficiency of processing the cases in the case scenario is improved, and the technical problem that the efficiency of processing the cases in the case scenario is low due to the fact that the case result and the used legal information cannot be predicted in the case scenario is solved.

Example 2

The technical solution of the present invention will be described below by way of example with reference to preferred embodiments.

The embodiment can be applied to research of integrated convenience service technology and equipment for litigation whole flow, aims to support cross-network system network mediation technology of multi-party access, and aims to improve mediation efficiency of human mediation personnel and judges by using artificial intelligence technology.

The embodiment provides a scheme for jointly predicting the litigation prediction information and the legal provision according to the case information submitted by the party, and the predicted litigation prediction information and the predicted legal provision can be used as reference results improved for a moderator, so that the moderator can conveniently make a mediation strategy, and meanwhile, the method can be flexibly applied to other types of judicial scenes, such as lawyers and the party.

Fig. 5 is a schematic diagram of a system for processing text information according to an embodiment of the present invention. As shown in fig. 5, the text information processing system 50 may include: a text preprocessing and pre-training module 51 and a neural network multitask prediction module 52.

The text preprocessing and pre-training module 51 is used for preprocessing the text information such as appeal, answer, evidence and the like, so that the subsequent neural network multi-task prediction module 52 can be conveniently used. The telltale and the answer in the text information are the most important case analysis materials, and the evidence of the original notice submission in the text information is also provided. Text describing and dissuading cases is a very important source of features. The embodiment mainly extracts two types of features aiming at the text information, including the statistical features of words, such as TF-IDF, and also including vector features.

Optionally, the preprocessing step of this embodiment includes: performing word segmentation processing on the text information, wherein any mainstream word segmentation tool can be used and a legal dictionary needs to be added; part-of-speech tagging, i.e., tagging each participle with part-of-speech tags, such as verbs, nouns, conjunctions, helpers, etc.; entity identification, for example, entity identification such as a person name, a place name, and an organization name is performed, which is not limited herein; statistical characteristics of TF-TDF, TF and the like can be mined.

When the word vector pre-training is carried out, the word vectors can be respectively trained aiming at different cases, and text information such as appeal, answer forms and judgment books of different cases can be used, wherein the more the text information is, the better the text information is. The method comprises the steps of segmenting words of text information of different cases, training word vectors through segmented corpora, and adopting a CBOW model, a Skip-gram model, an ELMo model, a bert model and the like without any limitation. Alternatively, this embodiment uses an ELMo model.

The neural network multi-tasking prediction module of this embodiment can be introduced as follows.

In this embodiment, the legal terms and the appeal results are two highly related tasks, and thus, an algorithm strategy for multi-task learning is selected with the purpose of enabling the two tasks to share features, improving the results of the two tasks.

Fig. 6 is a schematic diagram of an architecture of a multitasking model according to an embodiment of the present invention, and fig. 7 is a schematic diagram of a network structure of a multitasking model according to an embodiment of the present invention. As shown in fig. 6, the input layer includes a law information encoding layer, a case text information encoding layer, and a discrete feature input layer, the intermediate layer includes a law encoding layer and a complaint result encoding layer, and the output layer includes a law prediction result and a complaint prediction result.

The french information coding layer in this embodiment corresponds to the french information coding layer (law info embedding) in fig. 7, and fig. 7 is a schematic diagram of a network structure of a multitask model according to an embodiment of the present invention. In the legal information coding layer, the formal text of the legal provision is segmented, then each word is converted into a pre-trained word vector, and all the legal provisions can be coded into a word vector sequence. The word vectors of the embodiment can be pre-trained by adopting an ELMo model, and the method has the advantages that compared with the prior fixed word vectors, the ELMo model considers context information, and different context word vectors are different, so that the problems of ambiguous words and the like can be solved.

The case text information coding layer corresponds to the case text information coding layer (anti-level encoding) in fig. 7. The embodiment can perform word segmentation on case text first and convert the case text into a word vector sequence. And then coding the word vector sequence through cudnn/gru + self-attribute to finally obtain case text codes.

The discrete feature input layer corresponds to the discrete feature input layer (discrete feature) in fig. 7. The embodiment adopts the multilayer perceptron network to perform coding conversion on the discrete features of the case, and can convert the discrete feature vectors into a continuous semantic space, thereby being more suitable for training a deep learning model.

The slice encoding layer corresponds to the slice encoding layer (law embedding) of fig. 7. Optionally, in this embodiment, the case text information encoding layer and the discrete feature input layer are subjected to fusion encoding through Self-attention and highway strategies to obtain an encoding result, and then the encoding result is spliced with the french information encoding layer to finally obtain the french encoding layer.

The appeal result coding layer corresponds to the appeal result coding layer (acu embedding) in fig. 7. The method comprises the following steps of constructing an attribute label classifier (attr classifier), namely a multi-label classifier, wherein labels refer to legal labels and can be constructed manually, and optionally, the attribute label classifier is constructed through an attention mechanism. Obtaining a legal label coding layer; the second step is to splice the four items of a legal label coding layer (attr embedding), a law result coding layer (law result embedding, obtained by the intermediate result of the law classifier), a discrete feature input layer (discrete embedding) and a case text information coding layer, and then obtain an appeal result coding layer (accu embedding) by the highway strategy fusion coding.

The output layer of the French output layer is obtained by predicting through a French encoding layer (law embedding) and a Softmax function, and the appeal result is obtained by predicting through an appeal result encoding layer (acu embedding) and a Softmax function. The two tasks of the law statement prediction result and the appeal prediction result have the same coding mode in an input layer, the law statement prediction result is used as input information of the appeal result, and training of a model is added in the last layer. The optimization goal of the model of the embodiment is to synthesize the loss functions of the two subtasks to obtain the final loss function, and further ensure the high correlation of the model result through the two modes.

According to the case processing method and the case processing system, the case prediction information and the legal information are jointly predicted for the submitted cases through the multi-task prediction model trained in advance, wherein the case prediction information and the legal information have high correlation, so that the purpose of assisting in case approval is achieved, the efficiency of processing the cases in the case is improved, and the technical problem that the efficiency of processing the cases in the case is low due to the fact that the case result and the used legal information cannot be predicted in the case is solved.

Example 3

According to an embodiment of the present invention, there is also provided a text information processing apparatus for implementing the text information processing method shown in fig. 2.

Fig. 8 is a schematic diagram of a text information processing apparatus according to an embodiment of the present invention. As shown in fig. 8, the text information processing device 80 includes: a first acquisition unit 81, a first processing unit 82, and a first output unit 83.

The first acquiring unit 81 is configured to acquire target text information, where the target text information is used to describe case information of a case in a litigation scene.

The first processing unit 82 is configured to perform prediction processing on target text information based on a multitask prediction model, and predict legal information and litigation prediction information of a case, where the multitask prediction model is obtained by training text information of case information used for describing case samples in a litigation scene and legal information samples of the case samples.

And a first output unit 83 for outputting legal information and litigation prediction information of a case.

It should be noted here that the first acquiring unit 81, the first processing unit 82 and the first output unit 83 correspond to steps S202 to S206 in embodiment 1, and the three modules are the same as the corresponding steps in the implementation example and application scenario, but are not limited to the disclosure in the first embodiment. It should be noted that the above modules may be operated in the computer terminal 10 provided in embodiment 1 as a part of the apparatus.

According to the embodiment of the present invention, there is provided another text information processing apparatus for implementing the text information processing method shown in fig. 3. Fig. 9 is a schematic view of another text information processing apparatus according to an embodiment of the present invention. As shown in fig. 9, the processing 90 of the text information includes: a first display unit 91 and a second display unit 92.

The first display unit 91 is configured to display target text information on the text display interface, where the target text information is used to describe case information of cases in litigation scenes.

The second display unit 92 is configured to display, on the text display interface, legal information and litigation prediction information of a case, where the legal information and the litigation prediction information of the case are results of prediction processing performed on target text information based on a multitask prediction model, and the multitask prediction model is obtained by training text information of case information describing case samples in a litigation scene and legal information samples of the case samples.

It should be noted here that the first display unit 91 and the second display unit 92 correspond to step S302 and step S304 of embodiment 1, and the two modules are the same as the corresponding steps in the implementation example and application scenarios, but are not limited to the disclosure of the first embodiment. It should be noted that the above modules may be operated in the computer terminal 10 provided in embodiment 1 as a part of the apparatus.

According to an embodiment of the present invention, there is further provided another case information processing apparatus for implementing the case information processing method shown in fig. 4. Fig. 10 is a schematic diagram of a case information processing apparatus according to an embodiment of the present invention. As shown in fig. 10, the case information processing apparatus 100 includes: a second acquisition unit 101, a second processing unit 102 and a second output unit 103.

A second obtaining unit 101, configured to obtain case information of a case, where the case information of the case includes at least one of: appeal text, answer form text, legal and legal text, evidence text, and decision text.

The second processing unit 102 is configured to perform prediction processing on case information of a case based on a multitask prediction model, and predict legal information and litigation prediction information of the case, where the multitask prediction model is a neural network model and is configured to perform encoding processing on different types of files in the case information of the case.

And a second output unit 103 for outputting legal information and litigation prediction information of the case.

Example 4

The embodiment of the invention can provide a mobile terminal which can be any one computer terminal device in a computer terminal group.

Optionally, in this embodiment, the mobile terminal may be located in at least one network device of a plurality of network devices of a computer network.

In this embodiment, the mobile terminal may execute the program code of the following steps in the processing method of the text information of the application program: acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.

Alternatively, fig. 11 is a block diagram of a mobile terminal according to an embodiment of the present invention. As shown in fig. 11, the mobile terminal a may include: one or more processors 112 (only one shown), a memory 114, and a transmission device 116.

The transmission device is connected with the processor and is used for acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; and a memory coupled to the processor for providing instructions to the processor for processing the following processing steps: predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.

The memory may be used to store software programs and modules, such as program instructions/modules corresponding to the text information processing method and apparatus in the embodiments of the present invention, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, implements the above-described text information processing method. The memory may include high speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory located remotely from the processor, which may be connected to the mobile terminal a via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The processor can also call the information stored in the memory and the application program through the transmission device to execute the following steps: before target text information is processed based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, training the text information of a legal information sample and a case sample to obtain a first task prediction model, wherein the multi-task prediction model comprises a first task prediction model used for predicting the legal information of the case; and training the legal information samples, the text information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the multi-task prediction model comprises the second task prediction model, and the second task prediction model is used for predicting litigation prediction information of the case.

Optionally, the processor further executes program code for: coding the legal information sample to obtain a first coding result; coding the text information of the case sample to obtain a second coding result; coding discrete characteristics of the text information of the case sample to obtain a third coding result; training through the first coding result, the second coding result and the third coding result to obtain a first task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the first task prediction model; and training through the first coding result, the second coding result, the third coding result and legal information output by the first task prediction model to obtain a second task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the second task prediction model.

Optionally, the processor further executes program code for: performing word segmentation on a legal information sample to obtain a first word vector sequence; the first word vector sequence is determined as a first encoding result.

Optionally, the processor further executes program code for: segmenting words of the text information of the case sample to obtain a second word vector sequence; and coding the second word vector sequence to obtain a second coding result.

Optionally, the processor further executes program code for: and coding the discrete features through a multilayer perceptron to obtain a third coding result with continuous semantics.

Optionally, the processor further executes program code for: performing fusion coding on the second coding result and the third coding result to obtain a fourth coding result; splicing the fourth coding result and the first coding result to obtain a fifth coding result; and training through a fifth coding result to obtain the first task prediction model, wherein the fifth coding result is a result of an intermediate layer of the first task prediction model.

Optionally, the processor further executes program code for: and determining an output layer of the first task prediction model according to the fifth coding result and the objective function to obtain the first task prediction model.

Optionally, the processor further executes program code for: classifying the first coding result through a pre-constructed classifier of the legal information sample to obtain a classification result; splicing the classification result, the intermediate classification result used for obtaining the classification result, the second coding result and the third coding result to obtain a splicing result; performing fusion coding on the splicing result to obtain a sixth coding result, wherein the sixth coding result is the result of the middle layer of the second task prediction model; and training through the sixth coding result and legal information output by the first task prediction model to obtain a second task prediction model.

Optionally, the processor further executes program code for: determining an output layer of the second task prediction model according to the sixth coding result and the objective function; and training an output layer of the second task prediction model through legal information output by the first task prediction model to obtain the second task prediction model.

Optionally, the processor further executes program code for: acquiring a first loss function corresponding to legal information output by a first task prediction model; acquiring a second loss function corresponding to an output layer of the second task prediction model; and training an output layer of the second task prediction model through the first loss function and the second loss function to obtain the second task prediction model.

Optionally, the processor further executes program code for: extracting the statistical characteristics of words and the vector characteristics of the words from the text information of case samples; training through legal information samples, the statistical characteristics of words and the vector characteristics of the words to obtain a first task prediction model; training through the legal information samples, the text information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through legal information samples, the statistical characteristics of the words, the vector characteristics of the words and the legal information output by the first task prediction model to obtain a second task prediction model.

Optionally, the mobile terminal may execute the program code of the following steps in the processing method of the text information of the application program: displaying target text information on a text display interface, wherein the target text information is used for describing case information of cases in litigation scenes; and displaying legal information and litigation prediction information of the case on a text display interface, wherein the legal information and the litigation prediction information of the case are results of prediction processing on target text information based on a multi-task prediction model, and the multi-task prediction model is obtained by training text information of case information for describing case samples in litigation scenes and legal information samples of the case samples.

Optionally, the mobile terminal may execute the program code of the following steps in the processing method of case information of an application program: acquiring case information of a case, wherein the case information of the case comprises at least one of the following: appeal shape text, answer shape text, law and regulation text, evidence text and judgment text; predicting case information of a case based on a multitask prediction model to obtain legal information and litigation prediction information of the case, wherein the multitask prediction model is a neural network model and is used for coding different types of files in the case information of the case; and outputting legal information and litigation prediction information of the case.

The processor can call the information and application program stored in the memory through the transmission device to execute the following steps: before the case information is predicted based on the multitask prediction model, and legal information and litigation prediction information of the case are predicted, the method further comprises the following steps: training through legal information samples of case samples and case information of the case samples to obtain a first task prediction model, wherein the multi-task prediction model comprises a first task prediction model, and the first task prediction model is used for predicting the legal information of the cases; and training the legal information samples, the case information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the multi-task prediction model comprises the second task prediction model, and the second task prediction model is used for predicting litigation prediction information of the case.

Optionally, the processor further executes program code for: coding the legal information sample to obtain a first coding result; coding case information of the case sample to obtain a second coding result; coding discrete characteristics of case information of the case sample to obtain a third coding result; training through the first coding result, the second coding result and the third coding result to obtain a first task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the first task prediction model; and training through the first coding result, the second coding result, the third coding result and legal information output by the first task prediction model to obtain a second task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the second task prediction model.

By adopting the embodiment of the invention, the text information processing method is provided, and the target text information is obtained and is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case. That is, the case prediction information and the legal information are jointly predicted through the multi-task prediction model trained in advance, so that the purpose of assisting the case examination is achieved, the efficiency of processing the case in the litigation scene is improved, and the technical problem that the efficiency of processing the case in the litigation scene is low because the litigation result of the case and the used legal information cannot be predicted in the litigation scene is solved.

It can be understood by those skilled in the art that the structure shown in fig. 11 is only an illustration, and the Mobile terminal may also be a terminal device such as a smart phone (e.g., an Android phone, an iOS phone, etc.), a tablet computer, a palmtop computer, a Mobile Internet Device (MID), a PAD, and the like. Fig. 11 is a diagram illustrating a structure of the mobile terminal. For example, mobile terminal a may also include more or fewer components (e.g., network interfaces, display devices, etc.) than shown in fig. 11, or have a different configuration than shown in fig. 11.

Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by a program instructing hardware associated with the terminal device, where the program may be stored in a computer-readable storage medium, and the storage medium may include: flash disks, Read-Only memories (ROMs), Random Access Memories (RAMs), magnetic or optical disks, and the like.

Example 5

The embodiment of the invention also provides a storage medium. Alternatively, in this embodiment, the storage medium may be configured to store program codes executed by the text information processing method provided in embodiment 1.

Optionally, in this embodiment, the storage medium may be located in any one of computer terminals in a computer terminal group in a computer network, or in any one of mobile terminals in a mobile terminal group.

Optionally, in this embodiment, the storage medium is configured to store program code for performing the following steps: acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.

Optionally, the storage medium is further arranged to store program code for performing the steps of: before the target text information is processed based on the multitask prediction model to obtain legal information and litigation prediction information of a case, the method further comprises the following steps: training through the text information of the legal information sample and the case sample to obtain a first task prediction model, wherein the multi-task prediction model comprises a first task prediction model, and the first task prediction model is used for predicting the legal information of the case; and training the legal information samples, the text information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the multi-task prediction model comprises the second task prediction model, and the second task prediction model is used for predicting litigation prediction information of the case.

Optionally, the storage medium is further arranged to store program code for performing the steps of: coding the legal information sample to obtain a first coding result; coding the text information of the case sample to obtain a second coding result; coding discrete characteristics of the text information of the case sample to obtain a third coding result; training through the first coding result, the second coding result and the third coding result to obtain a first task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the first task prediction model; : and training through the first coding result, the second coding result, the third coding result and legal information output by the first task prediction model to obtain a second task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the second task prediction model.

Optionally, the storage medium is further arranged to store program code for performing the steps of: performing word segmentation on a legal information sample to obtain a first word vector sequence; the first word vector sequence is determined as a first encoding result.

Optionally, the storage medium is further arranged to store program code for performing the steps of: segmenting words of the text information of the case sample to obtain a second word vector sequence; and coding the second word vector sequence to obtain a second coding result.

Optionally, the storage medium is further arranged to store program code for performing the steps of: and coding the discrete features through a multilayer perceptron to obtain a third coding result with continuous semantics.

Optionally, the storage medium is further arranged to store program code for performing the steps of: performing fusion coding on the second coding result and the third coding result to obtain a fourth coding result; splicing the fourth coding result and the first coding result to obtain a fifth coding result; and training through a fifth coding result to obtain the first task prediction model, wherein the fifth coding result is a result of an intermediate layer of the first task prediction model.

Optionally, the storage medium is further arranged to store program code for performing the steps of: and determining an output layer of the first task prediction model according to the fifth coding result and the objective function to obtain the first task prediction model.

Optionally, the storage medium is further arranged to store program code for performing the steps of: classifying the first coding result through a pre-constructed classifier of the legal information sample to obtain a classification result; splicing the classification result, the intermediate classification result used for obtaining the classification result, the second coding result and the third coding result to obtain a splicing result; performing fusion coding on the splicing result to obtain a sixth coding result, wherein the sixth coding result is the result of the middle layer of the second task prediction model; and training through the sixth coding result and legal information output by the first task prediction model to obtain a second task prediction model.

Optionally, the storage medium is further arranged to store program code for performing the steps of: determining an output layer of the second task prediction model according to the sixth coding result and the objective function; and training an output layer of the second task prediction model through legal information output by the first task prediction model to obtain the second task prediction model.

Optionally, the storage medium is further arranged to store program code for performing the steps of: acquiring a first loss function corresponding to legal information output by a first task prediction model; acquiring a second loss function corresponding to an output layer of the second task prediction model; and training an output layer of the second task prediction model through the first loss function and the second loss function to obtain the second task prediction model.

Optionally, the storage medium is further arranged to store program code for performing the steps of: extracting the statistical characteristics of words and the vector characteristics of the words from the text information of case samples; training through legal information samples, the statistical characteristics of words and the vector characteristics of the words to obtain a first task prediction model; training through the legal information samples, the text information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through legal information samples, the statistical characteristics of the words, the vector characteristics of the words and the legal information output by the first task prediction model to obtain a second task prediction model.

Optionally, in this embodiment, the storage medium is further configured to store program code for performing the following steps: displaying target text information on a text display interface, wherein the target text information is used for describing case information of cases in litigation scenes; and displaying legal information and litigation prediction information of the case on a text display interface, wherein the legal information and the litigation prediction information of the case are results of prediction processing on target text information based on a multi-task prediction model, and the multi-task prediction model is obtained by training text information of case information for describing case samples in litigation scenes and legal information samples of the case samples.

Optionally, in this embodiment, the storage medium is further configured to store program code for performing the following steps: acquiring case information of a case, wherein the case information of the case comprises at least one of the following: appeal shape text, answer shape text, law and regulation text, evidence text and judgment text; predicting case information of a case based on a multitask prediction model to obtain legal information and litigation prediction information of the case, wherein the multitask prediction model is a neural network model and is used for coding different types of files in the case information of the case; and outputting legal information and litigation prediction information of the case.

Optionally, the storage medium is further arranged to store program code for performing the steps of: before case information is predicted based on a multi-task prediction model and legal information and litigation prediction information of a case are obtained through prediction, training is carried out on case information samples and case information of the case samples to obtain a first task prediction model, wherein the multi-task prediction model comprises a first task prediction model, and the first task prediction model is used for predicting the legal information of the case; and training the legal information samples, the case information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the multi-task prediction model comprises the second task prediction model, and the second task prediction model is used for predicting litigation prediction information of the case.

Optionally, the storage medium is further arranged to store program code for performing the steps of: coding the legal information sample to obtain a first coding result; coding case information of the case sample to obtain a second coding result; coding discrete characteristics of case information of the case sample to obtain a third coding result; training through the first coding result, the second coding result and the third coding result to obtain a first task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the first task prediction model; training through the legal information sample, the case information of the case sample and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through the first coding result, the second coding result, the third coding result and legal information output by the first task prediction model to obtain a second task prediction model, wherein the first coding result, the second coding result and the third coding result are results of an input layer of the second task prediction model.

Example 6

The embodiment of the invention also provides a processor. Optionally, in this embodiment, the processor runs the program of the text information processing method provided in embodiment 1.

Optionally, in this embodiment, the processor is configured to execute the program code of the following steps: acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; predicting target text information based on a multi-task prediction model to obtain legal information and litigation prediction information of a case, wherein the multi-task prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.

Optionally, in this embodiment, the processor is further configured to execute the program code of: displaying target text information on a text display interface, wherein the target text information is used for describing case information of cases in litigation scenes; and displaying legal information and litigation prediction information of the case on a text display interface, wherein the legal information and the litigation prediction information of the case are results of prediction processing on target text information based on a multi-task prediction model, and the multi-task prediction model is obtained by training text information of case information for describing case samples in litigation scenes and legal information samples of the case samples.

Optionally, in this embodiment, the processor is further configured to execute the program code of: acquiring case information of a case, wherein the case information of the case comprises at least one of the following: appeal shape text, answer shape text, law and regulation text, evidence text and judgment text; predicting case information of a case based on a multitask prediction model to obtain legal information and litigation prediction information of the case, wherein the multitask prediction model is a neural network model and is used for coding different types of files in the case information of the case; and outputting legal information and litigation prediction information of the case.

The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.

In the above embodiments of the present invention, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the embodiments provided in the present application, it should be understood that the disclosed technology can be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one type of division of logical functions, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, units or modules, and may be in an electrical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims

1. A method for processing text information, comprising:

acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes;

predicting the target text information based on a multitask prediction model to obtain legal information and litigation prediction information of the case, wherein the multitask prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples;

and outputting legal information and litigation prediction information of the case.

2. The method of claim 1, wherein prior to processing the target text information based on a multi-tasking prediction model to obtain legal information and litigation prediction information for the case, the method further comprises:

training through the legal information samples and the text information of the case samples to obtain a first task prediction model, wherein the multi-task prediction model comprises the first task prediction model, and the first task prediction model is used for predicting the legal information of the case;

and training the legal information samples, the text information of the case samples and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the multi-task prediction model comprises the second task prediction model, and the second task prediction model is used for predicting litigation prediction information of the case.

3. The method of claim 2,

the method further comprises the following steps: coding the legal information sample to obtain a first coding result; coding the text information of the case sample to obtain a second coding result; coding discrete characteristics of the text information of the case sample to obtain a third coding result;

training through the legal information sample and the text information of the case sample to obtain a first task prediction model, wherein the first task prediction model comprises the following steps: training through the first encoding result, the second encoding result and the third encoding result to obtain the first task prediction model, wherein the first encoding result, the second encoding result and the third encoding result are results of an input layer of the first task prediction model;

training through the legal information sample, the text information of the case sample and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through the first encoding result, the second encoding result, the third encoding result and legal information output by the first task prediction model to obtain the second task prediction model, wherein the first encoding result, the second encoding result and the third encoding result are results of an input layer of the second task prediction model.

4. The method of claim 3, wherein encoding the legal information sample to obtain a first encoding result comprises:

performing word segmentation on the legal information sample to obtain a first word vector sequence;

determining the first word vector sequence as the first encoding result.

5. The method of claim 3, wherein encoding the text information of the case sample to obtain a second encoding result comprises:

segmenting words of the text information of the case sample to obtain a second word vector sequence;

and coding the second word vector sequence to obtain the second coding result.

6. The method of claim 3, wherein encoding discrete features of the case sample's text information to obtain a third encoding result comprises:

and coding the discrete features through a multilayer perceptron to obtain a third coding result with continuous semantics.

7. The method of claim 3, wherein training through the first encoding result, the second encoding result, and the third encoding result to obtain the first task prediction model comprises:

performing fusion coding on the second coding result and the third coding result to obtain a fourth coding result;

splicing the fourth coding result and the first coding result to obtain a fifth coding result;

and training through the fifth coding result to obtain the first task prediction model, wherein the fifth coding result is a result of an intermediate layer of the first task prediction model.

8. The method of claim 7, wherein training through the fifth encoding result to obtain the first task prediction model comprises:

and determining an output layer of the first task prediction model according to the fifth encoding result and the objective function to obtain the first task prediction model.

9. The method of claim 3, wherein training through the first encoding result, the second encoding result, the third encoding result and legal information output by the first task prediction model to obtain the second task prediction model comprises:

classifying the first coding result through a pre-constructed classifier of the legal information sample to obtain a classification result;

splicing the classification result, an intermediate classification result used for obtaining the classification result, the second coding result and the third coding result to obtain a splicing result;

performing fusion coding on the splicing result to obtain a sixth coding result, wherein the sixth coding result is a result of an intermediate layer of the second task prediction model;

and training through the sixth coding result and legal information output by the first task prediction model to obtain the second task prediction model.

10. The method of claim 9, wherein training through the sixth encoding result and legal information output by the first task prediction model to obtain the second task prediction model comprises:

determining an output layer of the second task prediction model according to the sixth encoding result and an objective function;

and training an output layer of the second task prediction model according to legal information output by the first task prediction model to obtain the second task prediction model.

11. The method of claim 10, wherein training an output layer of the second task prediction model with legal information output by the first task prediction model, and wherein obtaining the second task prediction model comprises:

acquiring a first loss function corresponding to legal information output by the first task prediction model;

acquiring a second loss function corresponding to an output layer of the second task prediction model;

and training an output layer of the second task prediction model through the first loss function and the second loss function to obtain the second task prediction model.

12. The method according to any one of claims 2 to 11,

the method further comprises the following steps: extracting the statistical characteristics of words and the vector characteristics of the words from the text information of the case sample;

training through the legal information sample and the text information of the case sample to obtain a first task prediction model, wherein the first task prediction model comprises the following steps: training through the legal information sample, the statistical characteristics of the words and the vector characteristics of the words to obtain the first task prediction model;

training through the legal information sample, the text information of the case sample and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through the legal information sample, the statistical characteristics of the words, the vector characteristics of the words and the legal information output by the first task prediction model to obtain the second task prediction model.

13. A method for processing text information, comprising:

displaying target text information on a text display interface, wherein the target text information is used for describing case information of cases in litigation scenes;

and displaying legal information and litigation prediction information of the case on the text display interface, wherein the legal information and the litigation prediction information of the case are the results of prediction processing on the target text information based on a multi-task prediction model, and the multi-task prediction model is obtained by training text information of case information for describing case samples in the litigation scene and legal information samples of the case samples.

14. A case information processing method is characterized by comprising the following steps:

acquiring case information of a case, wherein the case information of the case comprises at least one of the following: appeal shape text, answer shape text, law and regulation text, evidence text and judgment text;

predicting the case information of the case based on a multitask prediction model to obtain legal information and litigation prediction information of the case, wherein the multitask prediction model is a neural network model and is used for coding different types of files in the case information of the case;

15. The method of claim 14, wherein prior to the case information being subjected to a prediction process based on a multitasking prediction model to predict legal information and litigation prediction information for the case, the method further comprises:

training through legal information samples of case samples and case information of the case samples to obtain a first task prediction model, wherein the multi-task prediction model comprises the first task prediction model, and the first task prediction model is used for predicting the legal information of the cases;

and training through the legal information sample, the case information of the case sample and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the multi-task prediction model comprises the second task prediction model, and the second task prediction model is used for predicting the litigation prediction information of the case.

16. The method of claim 15,

the method further comprises the following steps: coding the legal information sample to obtain a first coding result; coding the case information of the case sample to obtain a second coding result; coding discrete characteristics of case information of the case sample to obtain a third coding result;

training through the legal information sample and the case information of the case sample to obtain a first task prediction model, wherein the first task prediction model comprises the following steps: training through the first encoding result, the second encoding result and the third encoding result to obtain the first task prediction model, wherein the first encoding result, the second encoding result and the third encoding result are results of an input layer of the first task prediction model;

training through the legal information sample, the case information of the case sample and the legal information output by the first task prediction model to obtain a second task prediction model, wherein the second task prediction model comprises the following steps: and training through the first encoding result, the second encoding result, the third encoding result and legal information output by the first task prediction model to obtain the second task prediction model, wherein the first encoding result, the second encoding result and the third encoding result are results of an input layer of the first task prediction model.

17. An apparatus for processing text information, comprising:

the first acquisition unit is used for acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes;

the first processing unit is used for carrying out prediction processing on the target text information based on a multitask prediction model to obtain legal information and litigation prediction information of the case, wherein the multitask prediction model is obtained by training text information of case samples in the litigation scene and the legal information samples of the case samples;

and the first output unit is used for outputting legal information and litigation prediction information of the case.

18. An apparatus for processing text information, comprising:

the system comprises a first display unit, a second display unit and a third display unit, wherein the first display unit is used for displaying target text information on a text display interface, and the target text information is used for describing case information of cases in litigation scenes;

and the second display unit is used for displaying the legal information and the litigation prediction information of the case on the text display interface, wherein the legal information and the litigation prediction information of the case are the result of prediction processing on the target text information based on a multitask prediction model, and the multitask prediction model is obtained by training the text information of the case information for describing the case sample in the litigation scene and the legal information sample of the case sample.

19. An apparatus for processing case information, comprising:

a second obtaining unit, configured to obtain case information of a case, where the case information of the case includes at least one of: appeal shape text, answer shape text, law and regulation text, evidence text and judgment text;

the second processing unit is used for predicting the case information of the case based on a multitask prediction model to obtain legal information and litigation prediction information of the case through prediction, wherein the multitask prediction model is a neural network model and is used for coding different types of files in the case information of the case;

and the second output unit is used for outputting legal information and litigation prediction information of the case.

20. A storage medium comprising a stored program, wherein the program, when executed, controls an apparatus on which the storage medium is located to perform the steps of:

21. A processor, wherein the processor is configured to execute a program, wherein the program executes to perform the following steps:

22. A mobile terminal, comprising:

a processor;

the transmission device is connected with the processor and is used for acquiring target text information, wherein the target text information is used for describing case information of cases in litigation scenes; and

a memory coupled to the processor for providing instructions to the processor for processing the following processing steps: predicting the target text information based on a multitask prediction model to obtain legal information and litigation prediction information of the case, wherein the multitask prediction model is obtained by training text information of case information used for describing case samples in litigation scenes and legal information samples of the case samples; and outputting legal information and litigation prediction information of the case.