CN118861385B - Automatic scene generation method based on matching and computer equipment - Google Patents
Automatic scene generation method based on matching and computer equipment Download PDFInfo
- Publication number
- CN118861385B CN118861385B CN202411341608.0A CN202411341608A CN118861385B CN 118861385 B CN118861385 B CN 118861385B CN 202411341608 A CN202411341608 A CN 202411341608A CN 118861385 B CN118861385 B CN 118861385B
- Authority
- CN
- China
- Prior art keywords
- scene
- matching
- devices
- list
- matched
- 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
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000012216 screening Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 description 9
- 238000001914 filtration Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000003058 natural language processing Methods 0.000 description 3
- 238000013075 data extraction Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/9035—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/90335—Query processing
- G06F16/90344—Query processing by using string matching techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/9038—Presentation of query results
-
- 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention belongs to the technical field of intelligent home, and particularly relates to an automatic scene generation method based on matching and a computer device, wherein the method comprises the following steps of matching in a scene database according to a target scene name set by a user to obtain a matched scene result; storing all the devices in the matched scene result into a matching list, generating a device list according to the devices in the target scene, if the matching list contains devices similar to one device in the device list, then considering that the device is successfully matched, otherwise, considering that the device is failed to be matched, regarding the successfully matched devices as the control operation of the devices in the matching list, regarding the successfully matched devices as the control operation of the devices, using the trained large language model to generate the control operation of the devices, and integrating the control operation of all the devices to generate the target scene. The accuracy of braking scene generation is improved, and meanwhile, control operation of all devices in a generation target scene is guaranteed, and omission is avoided.
Description
Technical Field
The invention belongs to the technical field of intelligent home, and particularly relates to an automatic scene generation method based on matching and computer equipment.
Background
Scene creation is one of important functions in the field of smart home, and currently, two general methods are manual setting and template configuration. The manual setting is troublesome, manual operation is needed by a user, and the use friendliness is poor. The template configuration is set according to related rules, is convenient and fast compared with manual operation, but can only be generated according to a fixed template, and has insufficient flexibility.
In order to solve the technical problems and improve the flexibility and convenience of intelligent home scene generation, the Chinese patent application with the application publication number of CN117687314A provides a method for generating an intelligent home control scene based on a large language model. However, the large language model is artificial intelligence but is not a person at all, the accuracy of the finally generated scene is still unsatisfactory, and the problem is difficult to solve from the large model itself.
Disclosure of Invention
The invention aims to provide an automatic scene generation method based on matching and a computer device, which are used for solving the technical problem that scene generation precision is insufficient when a large language model is used for automatic scene generation.
In order to solve the technical problems, the invention provides an automatic scene generation method based on matching, which comprises the following steps:
Storing all devices in the matched scene result into a matching list, generating a device list according to the devices in the target scene, and if the device similar to a certain device in the device list exists in the matching list, judging that the device is successfully matched, otherwise, judging that the device is failed to be matched;
For equipment which is successfully matched, the control operation of the equipment in the matched list which is matched with the equipment is used as the control operation of the equipment;
and integrating the control operations of all the devices to generate a target scene.
Further, the process of generating the device list according to the devices in the target scene includes:
1) Obtaining a device type set comprising all device types according to the devices in the matched scene result;
2) Storing all types of devices in the target scene in a device type set into a device list;
Correspondingly, the automatic scene generation method further comprises the step of considering that the matching is failed for the devices with the types not in the device type set in the target scene.
Further, the method for matching in the scene database according to the target scene names set by the user to obtain the matched scene result comprises the following steps:
And further screening the scene results with the matching degree meeting the requirements, and taking the further screened results as matched scene results.
Further, the method for further screening the obtained scene results is that the scene results meeting the requirements are ranked according to the use times, and the scene results with higher use times are selected as the further screening results.
Further, the method for matching in the scene database according to the target scene names set by the user to obtain the matched scene result comprises the following steps:
And matching in a scene database according to the target scene names set by the user, and obtaining a scene result with the matching degree meeting the requirement as a matched scene result.
Further, the method for matching the equipment in the matching list and the equipment list comprises the steps of constructing character strings for the equipment in the matching list and the equipment list according to room names and equipment names of the equipment, and then performing similarity matching for the equipment in the matching list and the equipment list according to the constructed character strings.
Further, the method for judging whether the matching degree meets the requirement is that the semantic similarity between the scene name of the matching result and the scene name generated by the target is calculated, if the semantic similarity is larger than a similarity threshold, the scene result meets the requirement, otherwise, the scene result does not meet the requirement.
The method for performing similarity matching on the matching list and the equipment in the equipment list according to the constructed character strings comprises the steps of performing text similarity matching on the character strings of the equipment in the matching list and the equipment list, and selecting a matching result with higher matching degree to perform semantic similarity matching.
Further, the method for integrating the control operations of all the devices to generate the target scene comprises the steps of combining the device control operations generated by matching with the device control operations generated by the large language model, filtering the device control operations which do not accord with the logic of the target scene, and taking the filtered control operations of all the devices as the result of automatic scene generation.
The invention provides an improved invention, which has the beneficial effects that the automatic scene generating method based on matching filters the equipment in the scene, uses database matching to generate the control operation of the equipment which can be matched through the database, and uses a large model to generate the control operation of the equipment which can not be matched through the database. Specifically, a device list and a matching list are generated firstly, the device list and the matching list are matched, for successfully matched devices, control operations are generated according to matching results, for non-successfully matched devices, a trained large language model is used for generating the control operations, the accuracy is high, the situation that errors are generated automatically by the large language model is less likely to occur, therefore, the control operations of the database matching generation devices are preferentially used, the accuracy of automatic scene generation is higher, the safety and reliability are improved, and for non-successfully matched devices, the control operations of the non-successfully matched devices are generated by using the large language model, and the control operations of all the devices are guaranteed not to be missed.
To solve the above technical problem, the present invention also provides a computer device, including a processor, a memory and a computer program stored on the memory, where the processor is configured to execute the computer program to implement the method steps according to the matching-based automatic scene generating method of the present invention.
The invention provides an improved invention, which has the beneficial effects that the automatic scene generating method based on matching filters the equipment in the scene, uses database matching to generate the control operation of the equipment which can be matched through the database, and uses a large model to generate the control operation of the equipment which can not be matched through the database. Specifically, a device list and a matching list are generated firstly, the device list and the matching list are matched, for successfully matched devices, control operations are generated according to matching results, for non-successfully matched devices, a trained large language model is used for generating the control operations, the accuracy is high, the situation that errors are generated automatically by the large language model is less likely to occur, therefore, the control operations of the database matching generation devices are preferentially used, the accuracy of automatic scene generation is higher, the safety and reliability are improved, and for non-successfully matched devices, the control operations of the non-successfully matched devices are generated by using the large language model, and the control operations of all the devices are guaranteed not to be missed.
Drawings
FIG. 1 is a flow chart of an automatic scene generation method based on matching and large language models of a method embodiment of the present invention.
Detailed Description
The invention relates to an automatic scene generating method based on matching and a computer device, which filters devices in the scene, for devices that can be generated by database matching, their control operations are generated using database matching, and for devices that cannot be generated using database matching, their control operations are generated using a large model. Specifically, a device list and a matching list are generated firstly, the device list and the matching list are matched, for successfully matched devices, control operations are generated according to matching results, for non-successfully matched devices, a trained large language model is used for generating the control operations, the accuracy is high, the situation that errors are generated automatically by the large language model is less likely to occur, therefore, the control operations of the database matching generation devices are preferentially used, the accuracy of automatic scene generation is higher, the safety and reliability are improved, and for non-successfully matched devices, the control operations of the non-successfully matched devices are generated by using the large language model, and the control operations of all the devices are guaranteed not to be missed.
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent.
Method embodiment:
Referring to fig. 1, the matching-based automatic scene generation method of the present invention includes the steps of:
Step one, the algorithm receives scene names and related information and requests semantic services.
The algorithm receives the scene name input by the user and other related information of the scene, and requests the semantic model in the model to carry out semantic service.
And secondly, carrying out semantic processing on the scene name by the semantic model according to the request.
And thirdly, searching historical scene data with similar semantics from a database. The specific method comprises the following steps:
According to scene names which the user wants to generate, searching scenes with similar semantics of the scene names which are created by other users from a database in a semantic matching mode, calculating the semantic similarity between the names of the scenes which the user wants to generate and the scene names in the database, and taking the first 10 results with the semantic similarity of a certain scene in the database larger than a similarity threshold as scene results with matching degree meeting requirements. In this embodiment, the first 10 results with the semantic similarity greater than the threshold are taken, and as specific results of other embodiments, the number of the first 10 results can be freely set, when the number of the results with the semantic similarity greater than the threshold is less than 10, all the results with the semantic similarity greater than the threshold are taken, and the judgment conditions meeting the requirements can also be set, for example, the semantic similarity is not required to be greater than the threshold any more, the similarity is ordered from high to low in series, and the first 10 results are the results meeting the requirements.
In this embodiment, the matching is performed by using the semantic similarity between the target scene name and the scene name in the database, so that the automatic scene generation precision is higher, and as other embodiments, other matching methods may be adopted, for example, matching may be performed according to the overlapping degree of the devices.
And sorting the retrieved scenes meeting the requirements according to the use times of the scenes, wherein the scene ranking with more use times is higher, and obtaining the scene with the use times ranking before (including the set use times ranking) the set use times ranking (third name) as a matched scene result. In this embodiment, further filtering is performed by using a manner of sorting according to the number of times of use, and as other embodiments, sorting filtering may be performed according to other parameters such as the last time of use, or even filtering may not be performed, and the result of matching by the scene name may be directly used as the final matched scene result.
And step four, performing equipment behavior matching from a historical database. The specific method comprises the following steps:
All devices in the matching list (match_list) are traversed to generate a device type set (match_type_set) that includes all device types in the matching list (match_list).
All devices in the room involved in the current user setting scenario are retrieved, and a device list (device_list) is generated. In this embodiment, devices whose types exist in a device type set (match_type_set) among all devices in a room involved in a user setting scene are stored in a device list (device_list), and devices whose types do not exist in a device type set (match_type_set) are stored in a large model generation device list (device_ llm _list), and devices in the large model generation device list (device_ llm _list) are devices whose control operations cannot be generated using matching, and can be regarded as devices whose matching fails.
In this embodiment, when the device list is generated, filtering is performed in advance, and devices existing in the device type set are stored in the device list, so that the data processing amount of subsequent matching can be saved. As a further embodiment, all devices in the room involved in the current user setting scenario may also be directly stored in the device list.
And constructing character strings for the device list (device_list) and the matching list (match_list) according to the rules of 'room name + device name', matching text similarity according to the two rules of Bi-gram and LCS sequentially, screening out matching results with the matching degree above a set threshold, and matching semantic similarity according to the language model mapping to obtain a final matching result. If the equipment which meets the requirement on the matching degree of a certain equipment in the equipment list exists in the matching list, the equipment in the equipment list is considered to be successfully matched, and the control operation of the corresponding equipment in the matching list is taken as the control operation of the equipment in the equipment list. And adding the equipment list equipment with the matching degree not meeting the requirement into a large model generation equipment list (device_ llm _list), and taking out the corresponding equipment control operation in the matching list (match_list) which meets the condition as the control operation of the equipment to be matched, thereby completing the matching part.
The invention firstly uses Bi-gram rules to divide the text into two continuous sequences of letters or phonemes, and performs approximate matching degree between the two texts. Sequence similarity alignment was then performed using LCS (Longest Common Subsequence) rules. The text similarity comparison task mainly searches for the longest common subsequence in the two texts. The method is particularly suitable for comparing texts with different lengths, can ignore some irrelevant parts and mainly searches for longer common parts.
Semantic similarity screening of language model mapping:
Text is converted into a vector representation using a language model such as BERT (Bidirectional Encoder Representations from Transformers) or GPT (GENERATIVE PRE-trained Transformer), etc., and then similarity between vectors is calculated to measure semantic similarity for further matching.
The models can better capture semantic relations between words and sentences through pre-training the context information of the learning language, and further evaluate the semantic similarity of texts.
And fifthly, requesting a scene generation service.
And step six, generating the operation of the equipment with failed matching by using the large private domain model (large language model). The specific method comprises the following steps:
the names and operability lists of all devices in the device list (device_ llm _list) are extracted and sent to the trained large language model.
And extracting a result generated by the large language model, matching the equipment with the corresponding operation, removing the wrong item, and thus completing the large language model part.
The training process of the large language model is as follows:
1. Generating training data containing knowledge (automatic scene generation example) according to historical data extraction;
2. Generating training data which does not contain knowledge (an automatic scene generation example) according to historical data extraction;
3. and fusing the two data to perform end-to-end training of the large model. This way, large models can be prevented from overfitting;
4. training is performed, and the weight_decay is increased in the training process, so that the overfitting probability of the model is further reduced.
In order to prevent the problem of overfitting of the trained large model, training data containing no knowledge (automatic scene generation example) is added to the training data of the present embodiment.
And step seven, integrating the results of the matching part and the large language model part.
And step eight, verification processing and screening.
The method includes the steps of screening the following items which do not accord with target scene logic, for example, repeatedly switching on and off equipment, executing operation which needs to be switched on, heating in summer and refrigerating in winter of an air conditioner, simultaneously switching on and executing different playing operation of a plurality of sound box equipment, and improving the brightness of lamplight under the condition that film watching is needed.
And step nine, combining the generated control operations of all the user equipment to obtain a scene matched with the target scene of the user, and outputting the scene.
A natural language processing model (Natural Language Processing, NLP) asks the user if the generated scene is executed.
Through the method steps, the automatic scene generating method based on matching preferentially uses a semantic matching mode to search the control operation of the equipment in the scene values with similar names in the database, and uses the control operation of the equipment as the control operation of the user, and uses a large language model to generate the control operation of the equipment which cannot be generated through the semantic matching due to the data in the database.
Computer device embodiment:
The computer device of the invention comprises a processor, a memory and a computer program stored on the memory, wherein the processor is used for executing the computer program to realize the steps of the automatic scene generating method based on matching. Specific steps, principles, beneficial effects and the like of the method have been described in detail in the method embodiment, and the embodiment is not described in detail.
The processor may be a microprocessor MCU, a programmable logic device FPGA, or other processing device. The memory may be a variety of memory such as RAM, ROM, etc.
Claims (10)
1. An automatic scene generation method based on matching is characterized by comprising the following steps:
Storing all devices in the matched scene result into a matching list, generating a device list according to the devices in the target scene, and if the device similar to a certain device in the device list exists in the matching list, judging that the device is successfully matched, otherwise, judging that the device is failed to be matched;
For equipment which is successfully matched, the control operation of the equipment in the matched list which is matched with the equipment is used as the control operation of the equipment;
and integrating the control operations of all the devices to generate a target scene.
2. The automatic scene generating method based on matching according to claim 1, wherein the process of generating the device list from the devices in the target scene comprises:
1) Obtaining a device type set comprising all device types according to the devices in the matched scene result;
2) Storing all types of devices in the target scene in a device type set into a device list;
Correspondingly, the automatic scene generation method further comprises the step of considering that the matching is failed for the devices with the types not in the device type set in the target scene.
3. The automatic scene generating method based on matching according to claim 1, wherein the matching is performed in the scene database according to the target scene names set by the user, and the method for obtaining the matching scene result comprises the following steps:
And further screening the scene results with the matching degree meeting the requirements, and taking the further screened results as matched scene results.
4. The automatic scene generating method based on matching according to claim 3, wherein the method for further screening the obtained scene results is to sort the scene results meeting the requirements according to the number of use times, and select the scene result with higher number of use times as the result of further screening.
5. The automatic scene generating method based on matching according to claim 1, wherein the matching is performed in the scene database according to the target scene names set by the user, and the method for obtaining the matching scene result comprises the following steps:
And matching in a scene database according to the target scene names set by the user, and obtaining a scene result with the matching degree meeting the requirement as a matched scene result.
6. The automatic scene generating method based on matching according to claim 1, wherein the matching method for the devices in the matching list and the device list is to construct character strings for the devices in the matching list and the device list according to room names and device names thereof, respectively, and then to perform similarity matching for the devices in the matching list and the device list according to the constructed character strings.
7. The method for automatically generating a scene based on matching according to any one of claims 3 to 5, wherein the method for judging whether the matching degree meets the requirement is to calculate the semantic similarity between the scene name of the matching result and the scene name generated by the target, if the semantic similarity is greater than a similarity threshold, the scene result meets the requirement, otherwise the scene result does not meet the requirement.
8. The automatic scene generating method based on matching according to claim 6, wherein the method for performing similarity matching on the matching list and the devices in the device list according to the constructed character strings is that text similarity matching is performed on the character strings of the matching list and the devices in the device list, and matching results with higher matching degree are selected for semantic similarity matching.
9. The method for generating an automatic scene based on matching according to any of claims 1-6, wherein the method for integrating control operations of all devices to generate a target scene is to combine the device control operations generated by matching with the device control operations generated by the large language model, then filter out the device control operations which do not conform to the logic of the target scene, and take the filtered control operations of all devices as the result of the automatic scene generation.
10. A computer device comprising a processor, a memory and a computer program stored on the memory, characterized in that the processor is adapted to execute the computer program to carry out the steps of the match-based automatic scene generation method according to any of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411341608.0A CN118861385B (en) | 2024-09-25 | 2024-09-25 | Automatic scene generation method based on matching and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411341608.0A CN118861385B (en) | 2024-09-25 | 2024-09-25 | Automatic scene generation method based on matching and computer equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118861385A CN118861385A (en) | 2024-10-29 |
| CN118861385B true CN118861385B (en) | 2024-12-27 |
Family
ID=93177538
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411341608.0A Active CN118861385B (en) | 2024-09-25 | 2024-09-25 | Automatic scene generation method based on matching and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118861385B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114020909A (en) * | 2021-11-03 | 2022-02-08 | 深圳康佳电子科技有限公司 | Scene-based smart home control method, device, equipment and storage medium |
| CN114167736A (en) * | 2020-10-12 | 2022-03-11 | 超级智慧家(上海)物联网科技有限公司 | Intelligent household scene generation method and device |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US12221122B2 (en) * | 2022-10-18 | 2025-02-11 | Gm Cruise Holdings Llc | Synthetic scene generation for autonomous vehicle testing |
| CN117687314A (en) * | 2023-12-29 | 2024-03-12 | 四川启睿克科技有限公司 | Method for generating intelligent home control scene based on large language model |
-
2024
- 2024-09-25 CN CN202411341608.0A patent/CN118861385B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114167736A (en) * | 2020-10-12 | 2022-03-11 | 超级智慧家(上海)物联网科技有限公司 | Intelligent household scene generation method and device |
| CN114020909A (en) * | 2021-11-03 | 2022-02-08 | 深圳康佳电子科技有限公司 | Scene-based smart home control method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118861385A (en) | 2024-10-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108829757B (en) | Intelligent service method, server and storage medium for chat robot | |
| CN109240901B (en) | Performance analysis method, performance analysis device, storage medium, and electronic apparatus | |
| CN102004792B (en) | Method and system for generating hot-searching word | |
| CN111353310A (en) | Named entity identification method and device based on artificial intelligence and electronic equipment | |
| CN108563773B (en) | Knowledge graph-based legal provision accurate search ordering method | |
| CN111522909B (en) | Voice interaction method and server | |
| CN109299245B (en) | Method and device for recalling knowledge points | |
| CN113127632A (en) | Text summarization method and device based on heterogeneous graph, storage medium and terminal | |
| CN113971222B (en) | A multi-modal composite coding image retrieval method and system | |
| CN110442702A (en) | Searching method, device, readable storage medium storing program for executing and electronic equipment | |
| CN118051635B (en) | Conversational image retrieval method and device based on large language model | |
| CN116431837B (en) | Document retrieval method and device based on large language model and graph network model | |
| CN118377783B (en) | SQL sentence generation method and device | |
| US11914635B2 (en) | Performing image search based on user input using neural networks | |
| CN110765348A (en) | Hot word recommendation method and device, electronic equipment and storage medium | |
| CN117827886A (en) | Method for converting natural sentence into SQL sentence based on large language model | |
| CN111813916B (en) | Intelligent question-answering method, device, computer equipment and medium | |
| WO2023130951A1 (en) | Speech sentence segmentation method and apparatus, electronic device, and storage medium | |
| CN118861385B (en) | Automatic scene generation method based on matching and computer equipment | |
| CN118013020B (en) | Patent query method and system for generating joint training based on retrieval | |
| CN114328823A (en) | Database natural language query method and device, electronic device, storage medium | |
| CN117851543A (en) | Training method of text emotion recognition model, emotion recognition method and device | |
| CN117992592A (en) | Information processing method and device, electronic equipment and storage medium | |
| CN116992283A (en) | Language model dynamic update method, device, electronic equipment and storage medium | |
| CN113886535A (en) | Knowledge graph-based question and answer method and device, storage medium and electronic equipment |
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 |