CN113255358B - Multi-label character relation automatic labeling method based on event remote supervision - Google Patents
Multi-label character relation automatic labeling method based on event remote supervision Download PDFInfo
- Publication number
- CN113255358B CN113255358B CN202110782641.7A CN202110782641A CN113255358B CN 113255358 B CN113255358 B CN 113255358B CN 202110782641 A CN202110782641 A CN 202110782641A CN 113255358 B CN113255358 B CN 113255358B
- Authority
- CN
- China
- Prior art keywords
- event
- sentence
- character
- relationship
- marking
- 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.)
- Active
Links
Images
Classifications
-
- 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
- G06F40/295—Named entity recognition
-
- 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
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a multi-label character relation automatic labeling method based on event remote supervision, which comprises the following steps: collecting events influencing the relationship of the people; making a corresponding event marking template; constructing an event template knowledge base; and carrying out data preprocessing; carrying out event annotation on the preprocessed sentences by using an event annotation template; carrying out character relation annotation; marking the character relation to obtain a result; calculating the credibility between the 'event' and the 'sentence'; calculating a sentence score; setting a threshold value, and discarding sentences with sentence scores lower than the threshold value; and obtaining the final character relation label. According to the invention, the event influencing the character relationship is obtained through the matching of the event template knowledge base, and the character multi-label relationship is automatically deduced according to a plurality of events, so that the problem of character relationship multi-label can be solved, the precision of multi-label character relationship labeling can be obviously improved, and the method has more excellent mobility.
Description
Technical Field
The invention relates to the technical field of natural language relationship extraction, in particular to a multi-label character relationship automatic labeling method based on event remote supervision.
Background
Character relation extraction based on a deep learning method is widely applied to natural language processing tasks. The deep learning approach is data driven, requiring more labeled data. However, the manual labeling method requires a lot of manpower and time, and the data scale cannot satisfy the needs of the deep learning method. Remote surveillance is one of the common methods for automatically labeling a character relationship data set. It relies on a knowledge base containing a large number of triples to match pairs of people entities in a sentence with pairs of entities in the knowledge base. However, the construction of the triple knowledge base is time-consuming and labor-consuming, the relationships of all entities need to be exhausted, and the portability of the constructed triple knowledge base is low. Thus, a new event set-based automatic annotation method, event remote supervision (ESDS), is proposed. The method replaces a triple-based knowledge base with a small event-based knowledge set, which greatly reduces the cost of automatic annotation and provides considerable portability. However, the existing ESDS method only focuses on labeling of a single person relationship label, and ignores the characteristic that part of the person relationship may have multiple labels.
Disclosure of Invention
Therefore, the technical problem to be solved by the present invention is to overcome the above-mentioned defects in the prior art, and to provide a method for automatically labeling relationships among multiple tagged persons based on event remote supervision.
Therefore, the multi-label person relation automatic labeling method based on event remote supervision specifically comprises the following steps:
s1: collecting events influencing the character relation according to the classification type of the character relation data set required to be constructed;
s2: according to the determined event, making a corresponding event marking template: [ event trigger word, event-corresponding character relationship ];
s3: repeating S2 to make event marking templates corresponding to all the character relations, and constructing an event template knowledge base;
s4: collecting original texts, and carrying out data preprocessing, wherein the format of each preprocessed sentence is as follows [ entity1, entity2, sentence ];
s5: carrying out event annotation on the preprocessed sentence by using an event annotation template, and giving a plurality of event labels to the sentence when the sentence contains a plurality of event trigger words;
s6: carrying out character relation labeling on the sentence subjected to the event labeling;
s7: the result form obtained by performing the character relation labeling is as follows: d = [ Entity1, Entity2, Sennce, Event, Beginning Event, Score, internal Relationship ], wherein Entity1 and Entity2 represent human entities in a Sentence; sence represents the annotated Sentence; the Event represents an Event between the characters in the sentence, and if the Event is null, the Event represents that the current sentence has no corresponding Event; the Beginning Event refers to an Event that an entity in a sentence occurs in the sentence, and the internal Relationship is a character Relationship label marked in the sentence; score represents the confidence between the event and sentence;
s8: calculating the confidence between the "event" and the "sentenceS 1Represents; if "Event" is "NA", thenS 10, confidence between "Beginning Event" and "sentenceS 2Represents; calculating a sentence scoreSSThe following are:
SS =（r 1 S 1 - r 2（d i - d 0 ）+r 3 S 2）/(r 1 + r 2 +r 3)
whereinr 1To representS 1The weight coefficient of (a) is,r 2 a weight coefficient representing the distance of the sentence from the Beginning Event,r 3 to representS 2The weight coefficient of (a) is,d i indicating the index position of the current sentence in the text,d 0 indicating the index position of the Beginning Event in the text,d i andd 0 the difference represents the distance between the sentence and the Beginning Event;
s9: a threshold value is setk 2 Scoring the sentenceSSBelow thresholdk 2 The sentence of (1) is discarded;
s10: and after the processing of S9, obtaining the final character relation label.
Preferably, the classification categories of the physical relationship data set in S1 include "spouse relationship", "parent-child relationship", "superior-inferior relationship", "friendship relationship";
the events affecting the person relationship in S1 include: a "marriage event" marking the beginning of a "couple relationship" and a "divorce event" marking the end of a "couple relationship".
Preferably, the original text in S4 includes: character biography style linguistic data, novel style linguistic data, and character biography and novel style mixed linguistic data;
and (3) extracting sentence segmentation and word segmentation and character entities from the original text by using an NLTP tool, and reserving sentences which are more than or equal to two character entities in the sentences.
Preferably, S5 includes:
s5.1: clustering the preprocessed sentences according to the entity pairs to obtain a sentence set of the same entity pairs;
s5.2: matching the words in the sentence set with the event template knowledge base constructed in the S3;
s5.3: when the sentence contains the trigger word of the corresponding event, the current sentence is considered to have the corresponding event label;
s5.4: the sentence format after event annotation is as follows: entity1, entity2, sentence, [ event 1, event 2 … ] ].
Preferably, S6 includes:
s6.1: when the sentence A after the event marking has a corresponding event, directly marking the character relation corresponding to the sentence according to the event marking template;
s6.2: and when the labeled sentence B has no corresponding event, labeling the character relationship according to the event in the context.
Preferably, the sentence can correspond to a plurality of character relationship labels, and the satisfaction is a co-occurrence relationship.
Preferably, the confidence level between the "event" and the "sentenceS 1And embedding a vector by adopting a BERT model training word and calculating by cosine similarity.
Preferably, the threshold valuek 2 Is 0.5.
Preferably, S11: and marking the corresponding event according to the final character relation in the S10, making a corresponding event marking template, and supplementing the event marking template into an event template knowledge base.
According to the multi-label character relation automatic labeling method based on event remote supervision, provided by the invention, the event influencing the character relation is obtained through the matching of the event template knowledge base, and the character multi-label relation is automatically deduced according to a plurality of events, so that the character relation multi-label problem can be solved, the precision of labeling the character relation of the multi-label can be obviously improved, and the migration is more excellent. Meanwhile, the scoring mechanism in the ESDS method is improved by combining the semantic features and the distance features, and the automatic annotation precision of the character relation is further improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic flow chart of a multi-tag automatic character relationship labeling method according to the present invention;
FIG. 2 is a schematic diagram of an event template repository provided by the present invention;
FIG. 3 is a schematic flow chart of automatically labeling a multi-label relationship of a person based on an event template knowledge base according to the present invention;
FIG. 4 is a flow chart of the sentence score mechanism constructed based on semantics and distance according to the present invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. 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.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
The embodiment provides a multi-tag person relationship automatic labeling method based on event remote supervision, as shown in fig. 1 to 4, the method includes the following steps:
s1: collecting events influencing the character relation according to the classification type of the character relation data set required to be constructed; classification categories of the character relationship data sets such as "spouse relationship", "parent-child relationship", "superior-subordinate relationship", and "friendship relationship"; events that affect people's relationships are: a "marriage event" marking the beginning of a "couple relationship" and a "divorce event" marking the end of a "couple relationship".
S2: according to the determined event, making a corresponding event marking template: [ event trigger word, event-corresponding person relationship ].
S3: and repeating the step S2 to manufacture event marking templates corresponding to all the character relations, and constructing an event template knowledge base.
S4: collecting original texts, such as character biography style linguistic data, novel style linguistic data and character biography and novel style mixed linguistic data; carrying out data preprocessing, wherein the format of each preprocessed sentence is as follows [ entity1, entity2, sentence ]; and (3) extracting sentence segmentation and word segmentation and character entities from the original text by using an NLTP tool, and reserving sentences which are more than or equal to two character entities in the sentences.
S5: carrying out event annotation on the preprocessed sentence by using an event annotation template, and giving a plurality of event labels to the sentence when the sentence contains a plurality of event trigger words; the method specifically comprises the following steps: s5.1: clustering the preprocessed sentences according to the entity pairs to obtain a sentence set of the same entity pairs; s5.2: matching the words in the sentence set with the event template knowledge base constructed in the S3; s5.3: when the sentence contains the trigger word of the corresponding event, the current sentence is considered to have the corresponding event label; s5.4: the sentence format after event annotation is as follows: entity1, entity2, sentence, [ event 1, event 2 … ] ].
S6: carrying out character relation labeling on the sentence subjected to the event labeling; the method specifically comprises the following steps: s6.1: when the sentence A after the event marking has a corresponding event, directly marking the character relation corresponding to the sentence according to the event marking template; s6.2: and when the labeled sentence B has no corresponding event, labeling the character relationship according to the event in the context. If the entity pair has a "marriage event" in the context of a sentence and a "divorce event" in the context of a sentence, sentence B corresponds to a couple relationship resulting from the "marriage event".
A plurality of relationship labels corresponding to sentences need to satisfy the relationship that can be co-occurred, such as: the "superior-inferior relation" and the "parent-child relation" are co-occupiable, and the "couple relation" are not co-occupiable.
S7: the result form obtained by performing the character relation labeling is as follows: d = [ Entity1, Entity2, Sennce, Event, Beginning Event, Score, internal Relationship ], wherein Entity1 and Entity2 represent human entities in a Sentence; sence represents the annotated Sentence; the Event represents an Event between the characters in the sentence, and if the Event is null, the Event represents that the current sentence has no corresponding Event; the Beginning Event refers to an Event that an entity in a sentence occurs in the sentence, and the internal Relationship is a character Relationship label marked in the sentence; score represents the confidence between the event and sentence.
S8: calculating the confidence between the "event" and the "sentenceS 1Representing, embedding a vector by adopting a BERT model training word and calculating by cosine similarity; if "Event" is "NA", thenS 10, confidence between "Beginning Event" and "sentenceS 2Representing, embedding a vector by adopting a BERT model training word and calculating by cosine similarity; calculating a sentence scoreSSThe following are:
SS =（r 1 S 1 - r 2（d i - d 0 ）+r 3 S 2）/(r 1 + r 2 +r 3)
whereinr 1To representS 1The weight coefficient of (a) is,r 2 a weight coefficient representing the distance of the sentence from the Beginning Event,r 3 to representS 2The weight coefficient of (a) is,d i indicating the index position of the current sentence in the text,d 0 indicating the index position of the Beginning Event in the text,d i andd 0 the difference represents the distance of the sentence from the Beginning Event.
S9: a threshold value is setk 2 Scoring the sentenceSSBelow thresholdk 2 The sentence of (1) is discarded; threshold valuek 2 Is 0.5.
S10: and after the processing of S9, obtaining the final character relation label.
S11: and marking the corresponding event according to the final character relation in the S10, making a corresponding event marking template, and supplementing the event marking template into an event template knowledge base.
In the embodiment, the person event template knowledge base is used for replacing the original triple knowledge base, so that the action of the person entity is weakened, and the portability of the method is improved. The problem of wrong labeling caused by the fact that the triples are insufficient in the traditional remote method is solved. In addition, the data volume of the person event template is only related to the classification type of the data set to be constructed, and is far lower than the data set of the triple in the traditional method, so that a large amount of manpower and material resources can be saved.
According to the method for automatically labeling the multi-label character relationship based on event remote supervision, the event influencing the character relationship is obtained through matching the event template knowledge base, and the character multi-label relationship is automatically deduced according to the events, so that the problem of character relationship multi-label can be solved, the precision of labeling the multi-label character relationship can be obviously improved, and the method has better mobility. Meanwhile, the scoring mechanism in the ESDS method is improved by combining the semantic features and the distance features, and the automatic annotation precision of the character relation is further improved.
It should be understood that the above examples are only for clarity of illustration and are not intended to limit the embodiments. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. And obvious variations or modifications therefrom are within the scope of the invention.
Claims (1)
1. A multi-label character relation automatic labeling method based on event remote supervision is characterized by comprising the following steps:
s1: collecting events influencing the character relation according to the classification type of the character relation data set required to be constructed;
s2: according to the determined event, making a corresponding event marking template: [ event trigger word, event-corresponding character relationship ];
s3: repeating S2 to make event marking templates corresponding to all the character relations, and constructing an event template knowledge base;
s4: collecting original texts, and carrying out data preprocessing, wherein the format of each preprocessed sentence is as follows [ entity1, entity2, sentence ];
s5: carrying out event annotation on the preprocessed sentence by using an event annotation template, and giving a plurality of event labels to the sentence when the sentence contains a plurality of event trigger words;
s6: carrying out character relation labeling on the sentence subjected to the event labeling;
s7: the result form obtained by performing the character relation labeling is as follows: d = [ Entity1, Entity2, Sennce, Event, Beginning Event, Score, internal Relationship ], wherein Entity1 and Entity2 represent human entities in a Sentence; sence represents the annotated Sentence; the Event represents an Event between the characters in the sentence, and if the Event is null, the Event represents that the current sentence has no corresponding Event; the Beginning Event refers to an Event that an entity in a sentence occurs in the sentence, and the internal Relationship is a character Relationship label marked in the sentence; score represents the confidence between the event and sentence;
s8: calculating the confidence between the "event" and the "sentenceS 1Represents; if "Event" is "NA", thenS 10, confidence between "Beginning Event" and "sentenceS 2Represents; calculating a sentence scoreSSThe following are:
SS =（r 1 S 1 - r 2（d i - d 0 ）+r 3 S 2）/(r 1 + r 2 +r 3)
whereinr 1To representS 1The weight coefficient of (a) is,r 2 a weight coefficient representing the distance of the sentence from the Beginning Event,r 3 to representS 2The weight coefficient of (a) is,d i indicating the index position of the current sentence in the text,d 0 indicating the index position of the Beginning Event in the text,d i andd 0 the difference represents the distance between the sentence and the Beginning Event;
s9: is set up withA threshold valuek 2 Scoring the sentenceSSBelow thresholdk 2 The sentence of (1) is discarded;
s10: after the processing of S9, obtaining the final character relation label;
the classification categories of the human relationship data sets in S1 include "couple relationship", "parent-child relationship", "superior-inferior relationship", and "friendship relationship";
the events affecting the person relationship in S1 include: a "marriage event" marking the beginning of a "couple relationship" and a "divorce event" marking the end of a "couple relationship";
the original text in S4 includes: character biography style linguistic data, novel style linguistic data, and character biography and novel style mixed linguistic data;
extracting sentence segmentation, word segmentation and character entities from the original text by using an NLTP tool, and reserving sentences more than or equal to two character entities in the sentences;
s5 includes:
s5.1: clustering the preprocessed sentences according to the entity pairs to obtain a sentence set of the same entity pairs;
s5.2: matching the words in the sentence set with the event template knowledge base constructed in the S3;
s5.3: when the sentence contains the trigger word of the corresponding event, the current sentence is considered to have the corresponding event label;
s5.4: the sentence format after event annotation is as follows: entity1, entity2, sentence, [ event 1, event 2 … ] ];
s6 includes:
s6.1: when the sentence A after the event marking has a corresponding event, directly marking the character relation corresponding to the sentence according to the event marking template;
s6.2: when the marked sentence B has no corresponding event, marking out the character relationship according to the event in the context;
a plurality of character relation labels corresponding to sentences meet the requirement of co-occurrence;
confidence level between "event" and "sentenceS 1Training using BERT modelEmbedding words into vectors and calculating through cosine similarity;
threshold valuek 2 Is 0.5;
s11: and marking the corresponding event according to the final character relation in the S10, making a corresponding event marking template, and supplementing the event marking template into an event template knowledge base.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110782641.7A CN113255358B (en) | 2021-07-12 | 2021-07-12 | Multi-label character relation automatic labeling method based on event remote supervision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110782641.7A CN113255358B (en) | 2021-07-12 | 2021-07-12 | Multi-label character relation automatic labeling method based on event remote supervision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113255358A CN113255358A (en) | 2021-08-13 |
| CN113255358B true CN113255358B (en) | 2021-09-17 |
Family
ID=77191111
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110782641.7A Active CN113255358B (en) | 2021-07-12 | 2021-07-12 | Multi-label character relation automatic labeling method based on event remote supervision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255358B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104035975A (en) * | 2014-05-23 | 2014-09-10 | 华东师范大学 | Method utilizing Chinese online resources for supervising extraction of character relations remotely |
| US10388272B1 (en) * | 2018-12-04 | 2019-08-20 | Sorenson Ip Holdings, Llc | Training speech recognition systems using word sequences |
| CN112015871A (en) * | 2020-10-30 | 2020-12-01 | 中南大学 | Automatic character relation labeling method based on event set remote supervision |
| CN112784602A (en) * | 2020-12-03 | 2021-05-11 | 南京理工大学 | News emotion entity extraction method based on remote supervision |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10133728B2 (en) * | 2015-03-20 | 2018-11-20 | Microsoft Technology Licensing, Llc | Semantic parsing for complex knowledge extraction |
-
2021
- 2021-07-12 CN CN202110782641.7A patent/CN113255358B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104035975A (en) * | 2014-05-23 | 2014-09-10 | 华东师范大学 | Method utilizing Chinese online resources for supervising extraction of character relations remotely |
| US10388272B1 (en) * | 2018-12-04 | 2019-08-20 | Sorenson Ip Holdings, Llc | Training speech recognition systems using word sequences |
| CN112015871A (en) * | 2020-10-30 | 2020-12-01 | 中南大学 | Automatic character relation labeling method based on event set remote supervision |
| CN112784602A (en) * | 2020-12-03 | 2021-05-11 | 南京理工大学 | News emotion entity extraction method based on remote supervision |
Non-Patent Citations (1)
| Title |
|---|
| 基于结构化知识监督的事件抽取研究;曾颖;《人工智能》;20181231;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113255358A (en) | 2021-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108959242B (en) | Target entity identification method and device based on part-of-speech characteristics of Chinese characters | |
| CN110442841B (en) | Resume identification method and device, computer equipment and storage medium | |
| CN110795543A (en) | Unstructured data extraction method and device based on deep learning and storage medium | |
| CN111274239B (en) | Test paper structuring processing method, device and equipment | |
| CN111191275A (en) | Sensitive data identification method, system and device | |
| CN112035675A (en) | Medical text labeling method, device, equipment and storage medium | |
| CN111274804A (en) | Case information extraction method based on named entity recognition | |
| CN109460725B (en) | Receipt consumption details content mergence and extracting method, equipment and storage medium | |
| CN112686263B (en) | Character recognition method, character recognition device, electronic equipment and storage medium | |
| CN110941720A (en) | Knowledge base-based specific personnel information error correction method | |
| CN114444507A (en) | Context parameter Chinese entity prediction method based on water environment knowledge map enhancement relationship | |
| CN112559770A (en) | Text data relation extraction method, device and equipment and readable storage medium | |
| CN114153978A (en) | Model training method, information extraction method, device, equipment and storage medium | |
| CN110704615A (en) | Internet financial non-dominant advertisement identification method and device | |
| CN112015871B (en) | Automatic character relation labeling method based on event set remote supervision | |
| CN112395392A (en) | Intention identification method and device and readable storage medium | |
| CN112580351A (en) | Machine-generated text detection method based on self-information loss compensation | |
| CN115618883A (en) | Business semantic recognition method and device | |
| CN109960707B (en) | College recruitment data acquisition method and system based on artificial intelligence | |
| CN116796726A (en) | Resume analysis method, resume analysis device, terminal equipment and medium | |
| CN111597302B (en) | Text event acquisition method and device, electronic equipment and storage medium | |
| CN117034948B (en) | Paragraph identification method, system and storage medium based on multi-feature self-adaptive fusion | |
| CN113255358B (en) | Multi-label character relation automatic labeling method based on event remote supervision | |
| CN112487817A (en) | Named entity recognition model training method, sample labeling method, device and equipment | |
| CN111611806A (en) | Semantic analysis method, system and equipment for knowledge-graph question answering |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |