CN116383360A - Method and system for detecting answer body fitness of psychological consultation chat robot - Google Patents
Method and system for detecting answer body fitness of psychological consultation chat robot Download PDFInfo
- Publication number
- CN116383360A CN116383360A CN202310371990.9A CN202310371990A CN116383360A CN 116383360 A CN116383360 A CN 116383360A CN 202310371990 A CN202310371990 A CN 202310371990A CN 116383360 A CN116383360 A CN 116383360A
- Authority
- CN
- China
- Prior art keywords
- user
- word vector
- intention
- chat robot
- answer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 239000013598 vector Substances 0.000 claims abstract description 64
- 238000012549 training Methods 0.000 claims abstract description 23
- 230000006870 function Effects 0.000 claims abstract description 20
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 14
- 238000013528 artificial neural network Methods 0.000 claims abstract description 10
- 230000000694 effects Effects 0.000 claims abstract description 6
- 238000012545 processing Methods 0.000 claims abstract description 6
- 230000006399 behavior Effects 0.000 claims description 32
- 230000015654 memory Effects 0.000 claims description 16
- 238000004458 analytical method Methods 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 8
- 238000001514 detection method Methods 0.000 claims description 7
- 230000003542 behavioural effect Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000003993 interaction Effects 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 5
- 230000006872 improvement Effects 0.000 abstract description 4
- 238000005457 optimization Methods 0.000 abstract description 2
- 208000027418 Wounds and injury Diseases 0.000 description 3
- 230000006378 damage Effects 0.000 description 3
- 208000014674 injury Diseases 0.000 description 3
- 208000019901 Anxiety disease Diseases 0.000 description 2
- 230000036506 anxiety Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 208000033830 Hot Flashes Diseases 0.000 description 1
- 206010060800 Hot flush Diseases 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008033 biological extinction Effects 0.000 description 1
- 230000006854 communication Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 230000007794 irritation Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
- G10L15/1822—Parsing for meaning understanding
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Computer Interaction (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Multimedia (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Acoustics & Sound (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Databases & Information Systems (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method and a system for detecting the answer body subsidence of a psychological consultation chat robot, which relate to the technical field of language processing and comprise the following steps: acquiring the current input content of a user; converting the current input content of the user into text information to acquire user intention; carrying out word vector coding on the user intention to obtain word vector representation; inputting word vector representation of user intention into a neural network for training to obtain behavior distribution; calculating the relevance of the word vector representation and the behavior distribution by using a scoring function to obtain a calculated score; and splicing word vector representation, behavior distribution and calculated scores, and training by using a convolutional neural network to obtain the maximum value, namely the body fit degree of the psychological consultation chat robot answer. The invention is beneficial to the follow-up optimization and improvement of the conversation and treatment effects of the chat robot and is also beneficial to the improvement of the user satisfaction degree by detecting the answer body subsidence of the psychological consultation chat robot.
Description
Technical Field
The invention relates to the technical field of language processing, in particular to a method and a system for detecting the answer body subsidence of a psychological consultation chat robot.
Background
In recent years, with the push of artificial intelligence research on hot flashes, intelligent chat robots are positioned as portals for various products and services in the future, and become an important research project for companies in various related fields.
Particularly, the psychological consultation chat robot which is proposed aiming at psychological problems aims at pointedly processing unhealthy psychological conditions such as anxiety and the like, allows users to interact with electronic equipment through natural language, so that the chat robot has great potential for solving the problems, the unhealthy psychological conditions such as anxiety and the like and inconvenience caused by serious shortages of the number of psychological treatment doctors can be effectively relieved through the chat robot, treatment cost is greatly reduced, and the opportunity of patients to receive treatment is remarkably increased.
However, the psychological consulting chat robots on the market at present do not consider the factor of the body paste degree of the answers, and only the individual chat robots detect the positive extinction of the dialogue process, so that some answers generated in the question and answer process are too sharp, psychological problems of sensitive users can be expanded, and further irritation and injury are caused to the users.
Therefore, how to propose a method and a system for detecting the answer body paste of a psychological consultation chat robot, by detecting the answer body paste of the psychological consultation chat robot, detecting the body paste of a text, improving the answer body paste threshold of the psychological consultation chat robot, reducing further injury to a patient, optimizing and improving the dialogue and treatment effect of the chat robot, and improving the user satisfaction is a problem which needs to be solved by those skilled in the art.
Disclosure of Invention
In view of the above, the invention provides a method and a system for detecting the answer body intensity of a psychological consulting chat robot, which are used for detecting the answer body intensity of the psychological consulting chat robot, improving the answer body intensity threshold of the psychological consulting chat robot and reducing the further injury to patients, and in order to achieve the above purposes, the invention adopts the following technical scheme:
a method for detecting the answer body subsidence of a psychological consulting chat robot comprises the following steps:
acquiring the current input content of a user;
converting the current input content of the user into text information to obtain user intention;
performing word vector coding on the user intention to obtain word vector representation of the user intention;
inputting word vector representations of user intentions into a long-short-term memory neural network for training to obtain behavior distribution of the user;
calculating the correlation between the word vector representation of the user intention and the behavior distribution of the user by using a scoring function to obtain a calculated score;
and splicing word vector representation of the user intention, behavior distribution of the user and the calculated score, and training by using a convolutional neural network to obtain the maximum value, namely the body-building degree of the psychological consultation chat robot answer.
Optionally, the current input content of the user includes text, voice or gesture, wherein when the input text of the user through the man-machine interaction interface is detected, the input text is directly analyzed, and the user intention is obtained; when the voice input by the user is detected, a voice input mode is started, the microphone acquires the voice of the user, a corresponding text is generated through analysis, and the intention of the user is acquired through text analysis; when the user input gesture is detected, a video input mode is started, the camera collects the user gesture, corresponding text is generated through analysis, and the user intention is obtained through text analysis.
Optionally, the specific operation of performing word vector encoding on the user intention is to obtain each word in the user intention, and perform word vector encoding on text data submitted by the user in a word embedding manner to obtain a word vector I (x) of the user intention.
Optionally, the specific operation of obtaining the behavior distribution of the user is that the obtained word vector I (x) of the user intention is input into a two-way long-short-term memory network for training to obtain a vector g i As a behavioral profile of the user.
Optionally, the specific step of obtaining the calculated score is:
based on the user intention x and the selected behavior distribution g of the k most relevant users i Constructing relevant memory parameters as [ x, g ] 1 ，g 2 ]；
Calculating the relevance of all candidate words and the relevant memory parameters by using a scoring function;
obtaining the most relevant result r=argmax ωεW S R ([x,g 1 ,g 2 ],ω)；
Wherein ω is a candidate word, W is a set of all candidate words in the database, S R Is a function of the calculated score;
scoring function S R The following conditions are satisfied:
s(x,y)＝Φ x (x) T U T UΦ y (y)；
where U is an n D matrix, where n is the dimension, D is the number of features, Φ x And phi is y The effect of (a) is to map from the original text to the D-dimensional feature space.
Optionally, the method includes that word vectors of user intention, behavior distribution of the user and calculated scores are spliced to form new vectors, the new vectors are input into a convolutional neural network for training, and a result is output.
Optionally, the convolutional neural network includes: the full-connection layer is a full-classification connector, data subjected to pooling layer operation are input into the full-connection classifier, and probabilities of different body fit degrees are obtained through softmax function calculation:
P i ＝P(y i /w)；
wherein w represents an input text sequence of the system; y is i Representing an ith class; p (P) i ＝P(y i And/w) represents the probability of the ith class for a given sequence.
Optionally, a system for detecting the answer body fitness of a psychological consulting chat robot includes:
the acquisition module is used for: the method comprises the steps of obtaining current input content of a user;
and a conversion module: the method comprises the steps of converting the current input content of a user into text information to obtain user intention; performing word vector coding on the user intention to obtain word vector representation of the user intention; inputting word vector representations of user intentions into a long-short-term memory neural network for training to obtain behavior distribution of the user;
and a scoring module: calculating the correlation between the word vector representation of the user intention and the behavior distribution of the user by using a scoring function to obtain a calculated score;
the processing module is used for: and splicing word vector representation of user intention, behavior distribution of the user and the calculated score, and training by using a convolutional neural network to obtain the maximum value, namely the body-building degree of the psychological consultation chat robot answer.
Compared with the prior art, the invention discloses the answer body subsidence detection method and system for the psychological consultation chat robot, which have the following beneficial effects:
the method and the device acquire the current input content of the user; converting the current input content of the user into text information to acquire user intention; the body fit degree of the text is detected through the psychological consultation chat robot answer body fit degree detection, and in the dialogue, some dialogues can be positive but not body fit or negative but body fit, and in this case, a common emotion analysis method cannot accurately judge whether an answer is body fit/body fit. None of the existing psychological consultation chat robots detect the answer body paste of the chat robot, insufficient body paste and even pungent and aggressive answers are often generated in the communication process with users, and the mind of the psychological consultation patients is more sensitive and fragile. The invention detects the answer body intensity of the psychological consultation chat robot, is beneficial to the follow-up optimization and improvement of the conversation and treatment effects of the chat robot, and is also beneficial to the improvement of the user satisfaction.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a method for detecting the degree of answer in a psychological consulting chat robot.
Fig. 2 is a schematic diagram of a system for detecting the degree of answer in a psychological consulting chat robot.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The embodiment of the invention discloses a method for detecting the answer body subsidence of a psychological consultation chat robot, which comprises the following steps:
acquiring the current input content of a user;
converting the current input content of the user into text information to obtain user intention;
performing word vector coding on the user intention to obtain word vector representation of the user intention;
inputting word vector representations of user intentions into a long-short-term memory neural network for training to obtain behavior distribution of the user;
calculating the correlation between the word vector representation of the user intention and the behavior distribution of the user by using a scoring function to obtain a calculated score;
and splicing word vector representation of user intention, behavior distribution of the user and the calculated score, and training by using a convolutional neural network to obtain the maximum value, namely the body-building degree of the psychological consultation chat robot answer.
Further, a method for detecting the answer body fitness of a psychological consulting chat robot comprises the following steps:
step 1: preprocessing the current input content of a user, mainly comprising the following steps:
in the information acquisition module, the current input of the user is text, voice or gesture;
when detecting that a user inputs a text through a search box of a man-machine interaction interface, directly analyzing the input text to acquire user intention;
when the voice input by the user is detected, a voice input mode is started, the microphone acquires the voice of the user, a corresponding text is generated through analysis, and the intention of the user is acquired through text analysis;
when the user input gesture is detected, a video input mode is started, the camera collects the user gesture, corresponding text is generated through analysis, and the user intention is obtained through text analysis.
Step 2, the specific operation of word vector coding on the user intention is that each word in the user intention is obtained, word vector coding is carried out on text data submitted by the user in a word embedding mode, and word vector I (x) of the user intention is obtained;
and taking each word in the user intention as main input of the system, encoding the user intention by adopting a word embedding mode to obtain a word embedding matrix of the text sequence, and taking the word embedding matrix as input of a network in the next step. Each word x in the text x i As the main input of the system, the data file submitted by the user is encoded by adopting a word embedding mode to obtain a word embedding matrix of a text sequence, namely x i The word embedding vector representing the i-th word in the sequence takes the word embedding matrix as the input to the network in the next step.
Step 3: content capable of answering the user's question is initially determined based on the existing chat content as a behavior distribution of the answer content. In particular implementations, the existing chat content may include content of user questions received by the intelligent chat robot, as well as content answered by the intelligent chat robot for the user questions. And inputting the coded text data into a long-short-term memory neural network LSTM for training, and screening sentences which can answer user questions to serve as the behavior distribution. Specifically, the word vector I (x) of the user intention is input into a two-way long-short-term memory network for training to obtain a vector g i As a behavioral profile of the user.
Furthermore, after the existing chat content is determined, sentences in the existing chat content can be encoded through a hierarchical neural network, then the encoded sentences are used as input, long-term and short-term memory neural network is used for training to train sentences which are larger than a preset threshold value, the sentences are equivalent to screening out content which can answer user questions, irrelevant answers are removed, and the content is used as the behavior distribution of the intelligent chat robot for answering content which is made by the intelligent chat robot for user questions. Therefore, the response content possibly made by the intelligent chat robot can be screened once by utilizing a sequence-to-sequence deep learning method, so that the situation that behaviors possibly occur in a dialogue scene are borderless is reduced, namely, the response content which is irrelevant to a user question or has low relevance is reduced.
Step 4: calculating the relevance of the word vector representation of the user intention and the behavior distribution of the user by using a scoring function to obtain a calculated score, wherein the specific steps of obtaining the calculated score are as follows:
based on the user intention x and the selected behavior distribution g of the k most relevant users i Constructing relevant memory parameters as [ x, g ] 1 ，g 2 ]；
Calculating the relevance of all candidate words and the relevant memory parameters by using a scoring function;
obtaining the most relevant result r=argmax ωεW S R ([x,g 1 ,g 2 ],ω)；
Wherein ω is a candidate word, W is a set of all candidate words in the database, S R Is a function of the calculated score;
scoring function S R The following conditions are satisfied:
s(x,y)＝Φ x (x) T U T UΦ y (y)；
where U is an n D matrix, n is the dimension, D is the number of features, Φ x And phi is y The effect of (a) is to map from the original text to the D-dimensional feature space.
Step 5: splicing word vector representation of user intention, behavior distribution of the user and the calculated score, training by using a convolutional neural network to obtain the maximum value, namely, the body paste degree of the psychological consultation chat robot answer, splicing the word vector of the user intention, the behavior distribution of the user and the calculated score to form a new vector, inputting the new vector into the convolutional neural network for training, and outputting a result;
the convolutional neural network includes: the full-connection layer is a full-classification connector, data subjected to pooling layer operation are input into the full-connection classifier, and probabilities of different body fit degrees are obtained through softmax function calculation:
P i ＝P(y i /w)；
where w represents the input to the systemA text sequence; y is i Representing an ith class; p (P) i ＝P(y i W) represents the probability of the ith class for a given sequence;
P(Y/w)＝softmax(W z z+b z )；
where w represents the input text sequence of the system, Y represents all classifications, Y i Representing the ith class, P (Y|w) is the conditional probability of all classes in a given sequence, P i ＝P(y i W) represents the conditional probability of the ith class for a given sequence; w (W) z ∈R r ×n ，b z ∈R r Is a weight matrix and bias vector, z is a context vector for all time steps, and r is all classification numbers;
is the predicted output class.
In an embodiment, a system for detecting answer body engagement of a psychological consulting chat robot includes:
the acquisition module is used for: the method comprises the steps of obtaining current input content of a user;
and a conversion module: the method comprises the steps of converting the current input content of a user into text information to obtain user intention; performing word vector coding on the user intention to obtain word vector representation of the user intention; inputting word vector representations of user intentions into a long-short-term memory neural network for training to obtain behavior distribution of the user;
and a scoring module: calculating the correlation between the word vector representation of the user intention and the behavior distribution of the user by using a scoring function to obtain a calculated score;
the processing module is used for: and splicing word vector representation of user intention, behavior distribution of the user and the calculated score, and training by using a convolutional neural network to obtain the maximum value, namely the body-building degree of the psychological consultation chat robot answer.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. A method for detecting the degree of answer body attachment of a psychological consulting chat robot, comprising:
acquiring the current input content of a user;
converting the current input content of the user into text information to obtain user intention;
performing word vector coding on the user intention to obtain word vector representation of the user intention;
inputting word vector representations of user intentions into a long-short-term memory neural network for training to obtain behavior distribution of the user;
calculating the correlation between the word vector representation of the user intention and the behavior distribution of the user by using a scoring function to obtain a calculated score;
and splicing word vector representation of the user intention, behavior distribution of the user and the calculated score, and training by using a convolutional neural network to obtain the maximum value, namely the body-building degree of the psychological consultation chat robot answer.
2. The answer body subsidence detection method of a psychological consulting chat robot according to claim 1, wherein the current input content of the user comprises text, voice or gesture, wherein when the input text of the user through a man-machine interaction interface is detected, the input text is directly analyzed to obtain the intention of the user; when the voice input by the user is detected, a voice input mode is started, the microphone acquires the voice of the user, a corresponding text is generated through analysis, and the intention of the user is acquired through text analysis; when the user input gesture is detected, a video input mode is started, the camera collects the user gesture, corresponding text is generated through analysis, and the user intention is obtained through text analysis.
3. The method for detecting the answer body subsidence of a psychological consulting chat robot according to claim 1, wherein the specific operation of word vector coding for the user intention is to obtain each word in the user intention, and word vector coding is performed on text data submitted by the user in a word embedding manner to obtain word vector I (x) of the user intention.
4. The answer body fitness detection method of psychological consulting chat robot according to claim 3, characterized in that the specific operation of obtaining the behavior distribution of the user is that the obtained word vector I (x) of the user's intention is input into a two-way long-short-term memory network for training to obtain a vector g i As a behavioral profile of the user.
5. The answer body fitness detection method of a psychological consulting chat robot according to claim 1, wherein the specific steps of obtaining the calculated score are:
based on the user's intention x and the selected behavior distribution g of the k most relevant users i Constructing relevant memory parameters as [ x, g ] 1 ，g 2 ]；
Calculating the relevance of all candidate words and the relevant memory parameters by using a scoring function;
obtaining the most relevant result r=argmax ωεW S R ([x,g 1 ,g 2 ],ω)；
Wherein ω is a candidate word, W is a set of all candidate words in the database, S R Is a function of the calculated score;
scoring function S R The following conditions are satisfied:
s(x,y)＝Φ x (x) T U T UΦ y (y)；
where U is an n D matrix, where n is the dimension, D is the number of features, Φ x And phi is y The effect of (a) is to map from the original text to the D-dimensional feature space.
6. The method for detecting the answer body fitness of the psychological consulting chat robot according to claim 1, comprising the steps of splicing word vectors of user intention, behavior distribution of the user and calculated scores to form new vectors, inputting the new vectors into a convolutional neural network for training, and outputting results.
7. The answer body fitness detection method of a psychological consulting chat robot of claim 6, wherein the convolutional neural network comprises: the full-connection layer is a full-classification connector, data subjected to pooling layer operation are input into the full-connection classifier, and probabilities of different body fit degrees are obtained through softmax function calculation:
P i ＝P(y i /w)；
wherein w represents an input text sequence of the system; y is i Representing an ith class; p (P) i ＝P(y i And/w) represents the probability of the ith class for a given sequence.
8. A answer body fitness detection system of a psychological consulting chat robot, comprising:
the acquisition module is used for: the method comprises the steps of obtaining current input content of a user;
and a conversion module: the method comprises the steps of converting the current input content of a user into text information to obtain user intention; performing word vector coding on the user intention to obtain word vector representation of the user intention; inputting word vector representations of user intentions into a long-short-term memory neural network for training to obtain behavior distribution of the user;
and a scoring module: calculating the correlation between the word vector representation of the user intention and the behavior distribution of the user by using a scoring function to obtain a calculated score;
the processing module is used for: and splicing word vector representation of user intention, behavior distribution of the user and the calculated score, and training by using a convolutional neural network to obtain the maximum value, namely the body-building degree of the psychological consultation chat robot answer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310371990.9A CN116383360A (en) | 2023-04-10 | 2023-04-10 | Method and system for detecting answer body fitness of psychological consultation chat robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310371990.9A CN116383360A (en) | 2023-04-10 | 2023-04-10 | Method and system for detecting answer body fitness of psychological consultation chat robot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116383360A true CN116383360A (en) | 2023-07-04 |
Family
ID=86969041
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310371990.9A Pending CN116383360A (en) | 2023-04-10 | 2023-04-10 | Method and system for detecting answer body fitness of psychological consultation chat robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116383360A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116860952A (en) * | 2023-09-04 | 2023-10-10 | 富璟科技（深圳）有限公司 | RPA intelligent response processing method and system based on artificial intelligence |
-
2023
- 2023-04-10 CN CN202310371990.9A patent/CN116383360A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116860952A (en) * | 2023-09-04 | 2023-10-10 | 富璟科技（深圳）有限公司 | RPA intelligent response processing method and system based on artificial intelligence |
| CN116860952B (en) * | 2023-09-04 | 2023-11-03 | 富璟科技（深圳）有限公司 | RPA intelligent response processing method and system based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111554268B (en) | Language identification method based on language model, text classification method and device | |
| CN110609891A (en) | Visual dialog generation method based on context awareness graph neural network | |
| Yu et al. | On the integration of grounding language and learning objects | |
| CN113094578B (en) | Deep learning-based content recommendation method, device, equipment and storage medium | |
| CN110347787B (en) | Interview method and device based on AI auxiliary interview scene and terminal equipment | |
| CN112131883B (en) | Language model training method, device, computer equipment and storage medium | |
| CN113704428B (en) | Intelligent inquiry method, intelligent inquiry device, electronic equipment and storage medium | |
| CN111984780A (en) | Multi-intention recognition model training method, multi-intention recognition method and related device | |
| CN114020906A (en) | Chinese medical text information matching method and system based on twin neural network | |
| CN110597968A (en) | Reply selection method and device | |
| CN111597341A (en) | Document level relation extraction method, device, equipment and storage medium | |
| CN112100212A (en) | Case scenario extraction method based on machine learning and rule matching | |
| CN112101044A (en) | Intention identification method and device and electronic equipment | |
| CN116383360A (en) | Method and system for detecting answer body fitness of psychological consultation chat robot | |
| CN117198468A (en) | Intervention scheme intelligent management system based on behavior recognition and data analysis | |
| CN106708950B (en) | Data processing method and device for intelligent robot self-learning system | |
| CN115064154A (en) | Method and device for generating mixed language voice recognition model | |
| CN111428468A (en) | Method, device, equipment and storage medium for predicting single sentence smoothness | |
| CN116522165B (en) | Public opinion text matching system and method based on twin structure | |
| CN117493491A (en) | Natural language processing method and system based on machine learning | |
| CN112307179A (en) | Text matching method, device, equipment and storage medium | |
| CN110929013A (en) | Image question-answer implementation method based on bottom-up entry and positioning information fusion | |
| CN114357166A (en) | Text classification method based on deep learning | |
| CN112287690A (en) | Sign language translation method based on conditional sentence generation and cross-modal rearrangement | |
| CN112256833B (en) | Mobile phone problem intelligent question answering method based on big data and AI algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |