30 results about "Ambiguity" patented technology

The embodiment of the invention discloses a code generation method based on a large model, and the method comprises the steps: carrying out the grammatical analysis of a business logic process described by a natural language, and obtaining atomic task units and a logic relation between the atomic task units; and then re-editing according to the logical relationship to obtain a secondary expression text, and generating a code through LLM. And the problem of subsequent code generation through LLM due to logic relation chaos caused by natural language description for a business logic process is avoided. In the syntactic analysis process, the simplest expression corresponding to the natural language can be determined, ambiguity of description of steps to be executed in the business logic process is reduced, secondary expression texts are obtained through re-editing, the influence of logic chaos possibly occurring in the business logic process can be avoided, and the business logic processing efficiency is improved. Therefore, the normalization of the input LLM text is not affected regardless of the writing degree of the business logic process, and the accuracy and reliability of code generation are improved.

System and method for using knowledge fabric based on knowledge ontology, designed for deep language understanding and mechanistic causal reasoning, and meta-knowledge repository for auditable question answering. The method includes receiving an input text from a user, building a knowledge graph that represents real world facts and associations in the form of contextually tagged and weighted knowledge propositions, in multiple knowledge domains. The knowledge graph in combination with causal path knowledge and metadata describing digital sources containing answers constitutes the knowledge fabric. The method includes resolving ambiguity and determining actual intent of the user for the input text, from a plurality of interpretations of intent for sentences using the knowledge graph in conjunction with logical inference to achieve deep natural language understanding. The method includes finding / delivering response to the input request as to why / how unknown factors resulted in known outcome, or what outcomes are likely given known causal factors.

The invention discloses a question clarification method and device based on entity and intention fuzziness analysis, equipment, a medium and a product, and relates to the field of intelligent questions.The method comprises the steps that a question statement input by a user is obtained; cyclically executing the following steps until the entity ambiguity and the intention ambiguity of the problem statement are both smaller than or equal to a preset threshold: analyzing the entity and the intention in the problem statement to obtain the entity ambiguity and the intention ambiguity of the problem statement; under the condition that the entity ambiguity and / or the intention ambiguity are / is greater than a preset threshold value, generating target clarification request content based on entity ambiguity corresponding to the entity ambiguity and / or intention ambiguity corresponding to the intention ambiguity by utilizing a large language model; sending target clarification request content to the user, and obtaining target clarification response content of the user; and on the basis of the target clarification response content, the semantic representation of the question statement is updated, and the semantic ambiguity of the user question can be dynamically quantified by calculating the entity ambiguity and the intention ambiguity.

The invention discloses an NL2SQL ambiguity elimination method based on a large language model and rules in the technical field of natural language processing and database query. The NL2SQL ambiguity elimination method based on the large language model and the rule comprises the following steps: generating structured ambiguity mapping through the combination of the large language model and the rule, and covering a sentence pattern structure and pattern matching ambiguity; the ambiguity is clarified in the form of selection questions through user interaction and rule processing of a user interaction interface, and selection mapping is generated; and rewriting the natural language query and database mode based on the selection mapping, and outputting a final unambiguous problem and an unambiguous mode. According to the NL2SQL ambiguity elimination method based on the large language model and the rule, the accuracy and user friendliness of an analysis result are improved through a structured disambiguation framework, and clear SQL statements can be generated by inputting clear natural language query into an analyzer.

Techniques are disclosed herein for resolving date / time expressions while transforming natural language to a logical form such as a meaning representation language. A class label for a token in a natural language utterance and a meaning representation for the natural language utterance can be predicted. The class label can be associated with a date / time expression. The meaning representation can include an operator and a value. When the value associated with the class label matches a predetermined value type or the operator matches a predetermined operator, the value and / or the operator can be modified, and an executable statement can be generated for the meaning representation. A query on a computing system can be executed using the executable statement.

The invention discloses an intelligent report generation method and device, equipment, a storage medium and a program product, and relates to the technical field of data processing.The intelligent report generation method comprises the steps that a current natural language instruction input by a user is recognized, current operation information is obtained, and time sequence information in the current natural language instruction is obtained; searching in the historical operation information based on the time sequence information to obtain target time sequence operation information; based on matching of the target time sequence operation information and the current operation information, determining a current operation object corresponding to the current natural language instruction; and generating an intelligent report according to the operation information and the current operation object. As the historical operation information is matched and searched based on the time sequence information, the inaccuracy of report generation caused by operation ambiguity caused by fuzzy terms and lack of clear instructions is avoided; and matching constraint is performed through the current operation information and the target time sequence operation information, so that the understanding accuracy of the natural language instruction and the operation coherence are improved.

The application discloses a question clarification method, device, equipment, medium and product based on entity and intention ambiguity analysis, relates to the field of intelligent question answering, and comprises the following steps: acquiring a question sentence of a user input; the following steps are cyclically executed until the entity ambiguity and intention ambiguity of the question sentence are less than or equal to a preset threshold: analyzing entities and intentions in the question sentence to obtain the entity ambiguity and intention ambiguity of the question sentence; in the case where the entity ambiguity and / or intention ambiguity is greater than the preset threshold, generating target clarification request content by using a large language model based on entity ambiguity corresponding to entity ambiguity and / or intention ambiguity corresponding to intention ambiguity; sending the target clarification request content to the user and acquiring target clarification response content of the user; and updating the semantic representation of the question sentence based on the target clarification response content. The application can dynamically quantify the semantic ambiguity of a user question by calculating the entity ambiguity and intention ambiguity.

To provide an information processing system that can effectively improve work efficiency. [Solution] An information processing system S according to one embodiment includes an ambiguous expression detection and presentation unit (101, 201) that, when at least one of ambiguous terms and phrases is detected in part or all of a sentence contained in document data, presents at least one of the detected ambiguous terms and phrases, and a logic check unit (103, 203) that compares a sentence contained in one of a plurality of mutually associated document data with a sentence contained in the other document data, and presents a warning when a contradictory relationship is detected between a sentence contained in one of the plurality of document data and a sentence contained in the other document data.

The invention discloses a code annotation inconsistency detection method based on a large model and a related device thereof, code annotations of a target type in a smart contract are screened in combination with a binary classification method of mixed prompt, recognition can be well carried out according to a semantic form instead of a grammatical form, and then corresponding functions of the code annotations of the target type are extracted, so that code annotation inconsistency detection is realized. Obtaining source code fragments related to corresponding functions so as to carry out consistency identification on annotation-code pairs, and after fine adjustment is carried out on the annotation-code pairs which are input into a pre-training code generation model based on a consistency identification result, learning association of annotations and codes through the pre-training code generation model; the code annotation inconsistency detection result is obtained, the ambiguity problem is avoided, and the technical problems that according to an existing code annotation detection method, annotation texts cannot be accurately matched, and due to the fact that some annotated entities have semantic ambiguity, misjudgment is likely to happen when similarity analysis is conducted on the annotated entities and code facts are solved.

To provide a method and a system for resolving a hemispheric ambiguity using a position vector.SOLUTION: An embodiment resolves a hemispheric ambiguity in a system including a sensor. A handheld controller of the system emits a magnetic field. The sensor positioned in a headset of the system detects the magnetic field. A first position and an orientation of the handheld controller are determined in a first hemisphere to the headset based on the magnetic field. A second position and an orientation of the handheld controller are determined in a second hemisphere facing the first hemisphere to the headset in a diameter direction based on the magnetic field. A normal vector is determined to the headset, and a position vector identifies a position of the handheld controller to the headset in the first hemisphere.SELECTED DRAWING: Figure 25B

Term ambiguity is resolved by referencing a terminology database. An input is received comprising the term designated as ambiguous, and a string including the term. The term is posed as a query to the terminology database containing metadata of at least one type. Query results are returned including at least two possible meanings. Sequence(s) are extracted from the query results, each sequence including at least two pieces of metadata of a same type—one for each possible meaning of the ambiguous term. The metadata of each entry of a sequence is compared with the query result and corresponding scores are calculated. The scores are compared to determine a final meaning of the ambiguous term. Simpler embodiments considering one type of metadata (one sequence), may calculate and compare a listing of scores. Complex embodiments considering more than one type of metadata (multiple sequences), may calculate and compare a matrix of scores.

Techniques are described for determining one or more responses associated with one or more components in response to a user input. A system receives user input and causes one or more components to generate one or more responses associated with the user input. The system determines that one or more of the responses are responsive to the user input, causes one or more actions associated with the responses to be performed, and outputs a natural language summary of the one or more responses. If the system determines that no one of the responses is responsive to the user input and / or ambiguous with respect to the user input, the system can generate a request for additional information that can be used to address the ambiguity, the request for additional information can be sent to another component of the system and / or output to a user providing the user input.

To provide an information processing system capable of effectively improving work efficiency.SOLUTION: An information processing system S includes: ambiguous expression detection and presentation units (101 and 201) that, when at least one of a term and a phrase with ambiguity is detected in a part or all of text included in document data, present at least one of the detected term and phrase with ambiguity; and logical check units (103 and 203) that compare text included in one of a plurality of document data associated with each other with text included in other document data, and issue a warning when a contradictory relation is detected between the text included in the one of the plurality of document data and the text included in the other document data.SELECTED DRAWING: Figure 2

A method for ambiguity concept resolution for natural language processing includes receiving a message from a user; identifying an ambiguous concept in the message; determining a plurality of candidate concepts corresponding to the ambiguous concept; determining, for each of the plurality of candidate concepts, a respective similarity score based on a user-specific concept metric corresponding to the user. The method further includes ordering the plurality of candidate concepts based on the respective similarity scores; and determining that the ambiguous concept corresponds to a highest-ranked candidate concept of the ranked plurality of candidate concepts.

Methods, systems, and apparatus, including computer-readable media, for using artificial intelligence models to detect and mitigate ambiguity in natural language text. In some implementations, a system receives a text input from a user. The system classifies the text input as containing an ambiguity. Based on classifying the text input as containing an ambiguity, the system sends (i) a response to the text input generated using a first interpretation or meaning and (ii) one or more alternative interpretations or meanings. The system receives data indicating user input that selects one of the one or more alternative interpretations, and in response the system sends a second response generated using the selected alternative interpretation or meaning.

The present invention discloses a multi-round human-computer dialogue method and device based on entity ambiguity recognition and classification recommendation. The method and device obtain user input sentences; perform intent diagnosis on the input sentences; reply with relevant text when the diagnosis result is a chat-type intention; rewrite the input sentence containing ambiguous information when the diagnosis result is a knowledge dialogue-type intention, perform entity recognition and search to obtain candidate answers, perform semantic similarity judgment on the candidate answers based on the ambiguous information, and output the final answer. When the diagnosis result is a multi-round dialogue-type intention, the current input sentence is combined with the input sentence of the previous round of historical dialogue to perform a splicing operation, search for candidate answers on the spliced ​​sentence; perform semantic similarity judgment on the candidate answers, and output the final answer. The present invention solves the problem of ambiguous named entities being difficult to identify and having poor recognition in traditional human-computer dialogues, and the answers obtained are more accurate and the information in the knowledge graph is more fully utilized.

To make it easier for users to understand what factors are causing the text to be ambiguous. [Solution] The system includes an acquisition unit 141 that acquires target string data containing the target string to be checked for ambiguity, an acceptance unit 142 that accepts the specification of a threshold for determining whether or not the target string is ambiguous for each of the multiple types of ambiguity, an identification unit 143 that identifies the degree of ambiguity of the target string for each of the multiple types, and an output unit 145 that outputs information indicating that the target string is ambiguous for types where the degree of ambiguity is above the threshold.

This invention relates to the field of machine learning technology, specifically to a method and system for intelligent spatial data analysis and decision support based on a large language model. The method includes the following steps: acquiring spatial entities, location phrases, and task descriptions; analyzing word order and modification relationships; extracting task elements and action sequences; constructing continuous semantic paths; identifying geographic terms and range changes; extracting statements whose angle between action instruction direction vectors is less than a preset threshold; and obtaining a set of spatial action structures. In this invention, a semantic task list is constructed by extracting spatial entities, location phrases, and task content. Actions are ordered by combining action words and word order relationships. The connection methods between segments are analyzed, and a coherent task path is organized. The sequential matching and semantic connection between spatial objects and actions are completed. By filtering the information structure of action direction and location changes, task statements are arranged in an orderly manner according to the process, avoiding ambiguity and interruptions in instruction transmission, and maintaining the orderly integration and flow of spatial information.

Methods, systems, and non-transitory computer readable storage media are disclosed for detecting specific ambiguity types in queries to a large language model. The disclosed system determines a query in a prompt by a client device to a large language model. The disclosed system generates, utilizing a small language model, a label indicating that the query comprises an ambiguity of an identified ambiguity type of a plurality of ambiguity types according to a plurality of quantitative features of the query. Additionally, the disclosed system generates, for display to the client device, a response to the query based on the ambiguity of the identified ambiguity type.

To provide an information processing system capable of effectively improving work efficiency.SOLUTION: An information processing system S includes: ambiguous expression detection and presentation units (101 and 201) that, when at least one of a term and a phrase with ambiguity is detected in a part or all of text included in document data, present at least one of the detected term and phrase with ambiguity; and logical check units (103 and 203) that compare text included in one of a plurality of document data associated with each other with text included in other document data, and issue a warning when a contradictory relation is detected between the text included in the one of the plurality of document data and the text included in the other document data.SELECTED DRAWING: Figure 2

The application relates to the technical field of natural language processing, and in particular to a multi-stage retrieval enhancement generation method, a computer system and a computer readable storage medium. Whether external retrieval is needed for an input question is detected, ambiguity and decomposition are detected in sequence, in the ambiguity elimination stage, when the model determines that the question has ambiguity, the problem semantics is clarified by constructing a prompt word template, so that the error retrieval caused by ambiguity is avoided. In the question decomposition stage, corresponding decomposition strategies are adopted according to different question types to ensure the integrity and accuracy of the retrieval and reasoning process. Finally, in the answer generation stage, the model comprehensively utilizes the external retrieval content and the self-knowledge according to the discrimination result to generate the answer, and integrates the sub-question answers to obtain the final output result. The application aims to solve the problem of how to realize accurate semantic retrieval of complex questions while controlling the computational overhead of the multi-stage retrieval enhancement generation method.

A generative pre-trained large language model (LLM) may create domain-specific text answers in various formats such as JSON, XML, HTML, SQL, or programming language. However, LLM may "appear illusion", generating incorrect or meaningless answers that deviate from reality, thereby weakening trust on its output or causing more serious consequences. The disclosed techniques use a sampling-based method and an equivalence checker. Generating, by the LLM, a plurality of answers (samples) to the cue; if the plurality of answers are equivalent, the LLM is likely to answer correctly. If the samples are not consistent or contradictory, it is more likely that the LLM has illusion, or ambiguity is prompted. The functional equivalence of the sample is validated using an automatic inference equivalence verifier, which provides a method of detecting and possibly correcting hallucination questions in the LLM generated answers.

The present invention discloses a description and practice method for inferable and precise semantics based on a knowledge graph, which relates to the field of intelligent education technology. In the present invention, a cell description language for the mathematics subject is defined to describe the semantics of questions, and a complete description framework for concepts, individuals, and relationships is defined, which not only includes common binary relationships but also extends to the description of multi - element relationships. At the same time, for multi - element relationships, corresponding Atom extension rules are defined to obtain all atomic information from simple relationship descriptions, providing a basis for knowledge reasoning, eliminating the diversity and ambiguity of natural language, and giving definitions with uniqueness and accuracy. In addition, aiming at the problem of non - uniqueness of semantic information, the normalization rules of the cell language are defined to eliminate ambiguity, solving the problem that the existing methods cannot accurately and unambiguously describe the semantics of questions. On this basis, the present invention can realize the intelligent solution of questions through an inference engine, greatly reducing the teaching cost and tutoring cost of teachers.

The invention relates to the technical field of machine learning, in particular to a spatial data intelligent analysis decision support method and system based on a large language model.The method comprises the following steps that spatial entities, position phrases and task description are obtained, word order and modification relations are analyzed, task elements and action sequences are extracted, and a continuous semantic path is constructed; and identifying geographical words and range changes, and extracting statements with action instruction direction vector included angles smaller than a preset threshold value to obtain a spatial action structure set. According to the method, a semantic task list is constructed by extracting space entities, position phrases and task contents, action sorting is completed by combining action words and word order relations, connection modes between phrases are sorted, a coherent task path is organized and formed, and sequence matching and semantic connection between space objects and actions are completed. By screening information structures of action directions and position changes, task statements are pushed to be arranged in order according to a process mode, ambiguity and interruption in instruction transmission are avoided, and ordered fusion and circulation of spatial information are kept.

The present application relates to a kind of Chinese word sense disambiguation method of combining multi-channel hybrid dilated convolution neural network with residual and attention (Combining Multi-Channel Hybrid Dilated Convolution Neural Network with Residual and Attention, MHDCNN-RA). The present application first carries out word segmentation, part-of-speech tagging, semantic class labeling to Chinese sentence containing ambiguous words, obtains processed training corpus and test corpus. Then, the word sense disambiguation model is trained using the training corpus, and the optimized MHDCNN-RA model is obtained. On the optimized MHDCNN-RA model, the test corpus is disambiguated, and the weight of the ambiguous word under each semantic category is obtained. The semantic category with the maximum weight is the semantic category of the ambiguous word. The present application realizes good disambiguation of ambiguous words and more accurately judges the true meaning of ambiguous words.

The present disclosure relates to a device for resolving textual ambiguity through a visual language inference model, wherein the device includes: a pun identification unit that input an image and an original text, identifies a pun phrase in the original text utilizing the image as a clue, and generates a plurality of candidate translations; a pun semantic interpretation unit that input the plurality of candidate pun translations into the visual language inference model and decides a pun translation based on consistency with the image; and a pun reconstruction unit that reconstructs the pun translation by reflecting an intention of the original text.

The present invention relates to the field of digital technology, and in particular to a digital interactive learning model system and an operation method. In the present invention, a digital interactive learning model is tested, a test question sentence is input into the digital interactive learning model, an output reply sentence is matched with a test reply sentence corresponding to a pre-stored test question sentence, features in the reply sentence and the test reply sentence are respectively extracted, the operation of the model is analyzed based on the number of identical features, the ambiguity of the test question sentence is analyzed according to the number of associated words of each feature in the test question sentence, and the degree of influence of the test question sentence on the operation of the model is analyzed according to the ambiguity of the test question sentence. The more the number of associated words, the more interpretations of the question sentence, and the greater the influence on the operation of the model. The operation of the model is analyzed according to the number of associated words. When there are many associated words, the number of subsequent reply sentences is increased to improve the interactive learning efficiency of the model.

The invention relates to the technical field of natural language processing, in particular to a multi-stage retrieval enhancement generation method, a computer system and a computer readable storage medium. According to the method, whether external retrieval, ambiguity and decomposition detection are needed or not is conducted on an input problem in sequence, in the ambiguity eliminating stage, when a model judges that ambiguity exists in the problem, definition of semantic meaning of the problem is achieved by constructing a cue word template, and therefore wrong retrieval caused by ambiguity is avoided. In the problem decomposition stage, corresponding decomposition strategies are adopted according to different problem types, so that the completeness and accuracy of the retrieval and reasoning process are guaranteed. And finally, in an answer generation stage, the model comprehensively utilizes external retrieval content and self knowledge to generate answers according to a judgment result, and the answers of the sub-questions are integrated to obtain a final output result. The objective of the invention is to solve the problem of how to control the calculation overhead of a multi-stage retrieval enhancement generation method while realizing accurate semantic retrieval of complex problems.

The application relates to the technical field of data retrieval, in particular to a method for determining an ambiguous word, a retrieval method and an electronic device. The method for determining the ambiguous word comprises the following steps: obtaining an initial subject word; based on the initial subject word, performing segmentation and splicing on a preset text to obtain a target subject word containing the initial subject word; wherein the target subject word is segmented in a first segmentation mode to obtain a plurality of words containing the initial subject word, and is segmented in a second segmentation mode to obtain a plurality of continuous non-initial subject words; and if the number of times that the target subject word appears in the preset text is greater than a preset value, the target subject word is determined as an ambiguous word. In this way, based on the scheme provided in the application, the ambiguous word of the initial subject word can be automatically identified, and compared with the manual determination scheme in the background art, the scheme provided in the application is more time-saving and labor-saving.

The invention relates to an output method and system for automatically generating a cue word optimization large model based on user intention, and the method comprises the following steps: (1) receiving natural language input, and generating a preliminary response through a main large language model; (2) analyzing historical dialogues and current input in real time, and cooperatively judging whether a user intention is clear or not by adopting multiple mechanisms; (3) when the intention is determined to be clear, generating an optimization cue word based on the dialogue context; (4) inputting the optimized cue word into the main large language model to generate a final response; and (5) outputting a final response. The method has the advantages that the real intention of the user is automatically recognized, and dialogue ambiguity and misleading information are reduced through multi-mechanism intention judgment; prompt word optimization is automatic, and the problem that a user manually summarizes or rewrites is avoided.