54 results about "Sentence similarity" patented technology

Systems and methods are described herein for detecting and reducing hallucinations and improving the reliability of an LLM-based QA system. Particularly, the disclosure provides a set of hallucination detection and handling approaches that address different types of hallucinations of the QA system. The hallucination detection approaches described herein include comparing a natural language answer with context information by way of sentence similarity estimation and keyword matching. The hallucination handling approaches described herein include removing hallucinated sentences from the natural language answer or regenerating the natural language answer using better context information, depending on a level of hallucination detected in the natural language answer. A hybrid framework is also provided that systematically combines the hallucination detection approaches and hallucination handling approaches into one system to achieve an optimal hallucination-reduction performance for the QA system.

Method and apparatus for dynamic remediation. A set of records associated with a project is accessed. The set of records is processed using one or more natural language processing techniques to generate textual data comprising a plurality of pairs of sentences corresponding to one or more topics associated with the project. An issue for at least one topic associated with the project is identified based on the textual data, comprising identifying a pair of sentences that comprises a first sentence and a second sentence, calculating a sentence similarity score by comparing the first and second sentences using a similarity metric, and determining that the sentence similarity score satisfies one or more criteria. In response to determining that the one or more criteria are satisfied, a project meeting for the issue is scheduled based at least in part on a criticality of the issue.

The invention provides a law decision prediction method and device based on a large language model and logic enhancement, which can realize adaptive adjustment of law decision rules and improve the accuracy of decision prediction, and comprises the following steps: generating decision rules in the form of a first-order logic language by adopting an LLM model in combination with a law article original text and a historical case, describing key elements in legal provisions and historical cases through first-order logic symbols to form a consequent of a judgment rule, and taking a judgment label as a consequent of the judgment rule; calculating statement similarity of the cases based on a BGE vector model, and screening historical cases with high similarity to construct a confusion-prone case verification set; based on the confusion-prone case verification set, using a confusion perception comparative learning method to optimize a judgment rule; and for a to-be-predicted target case, generating a candidate tag of the target case based on a pre-trained domain small model, judging whether the target case accords with the candidate tag through the LLM model in combination with the thinking chain and the optimized judgment rule, and outputting a predicted judgment result.

The invention discloses a Tibetan sentence similarity calculation method based on feature fusion. The Tibetan sentence similarity calculation method comprises the following steps of S1, performing boundary recognition on Tibetan sentences; s2, preprocessing the Tibetan sentences recognized in the step S1, including distinguishing the sentences according to minimum semantic unit word granularity, and expressing the sentences by using trained word vectors to obtain a matrix form of the sentences; s3, constructing a sentence representation model to analyze the sentences to obtain sentence vectors; s4, calculating a similarity value based on the obtained sentence vector by using a sentence similarity calculation model; and S5, outputting the similarity. Compared with the prior art, the Tibetan sentence similarity calculation method based on the feature fusion has the advantage that the Tibetan sentence similarity calculation method based on the Tibetan application is provided.

Method and apparatus for dynamic remediation. A set of records associated with a project is accessed. The set of records is processed using one or more natural language processing techniques to generate textual data comprising a plurality of pairs of sentences corresponding to one or more topics associated with the project. An issue for at least one topic associated with the project is identified based on the textual data, comprising identifying a pair of sentences that comprises a first sentence and a second sentence, calculating a sentence similarity score by comparing the first and second sentences using a similarity metric, and determining that the sentence similarity score satisfies one or more criteria. In response to determining that the one or more criteria are satisfied, a project meeting for the issue is scheduled based at least in part on a criticality of the issue.

The invention relates to a patient follow-up visit management method and system, and solves the problems that generation of follow-up visit content by means of a fixed template is lack of personalized consideration, and actual situations such as special requirements, living changes and current disease treatment stages of patients cannot be fully combined. The method comprises the following steps: performing feature coding on a first draft of follow-up content and personalized adjustment information, identifying semantic association features of sentence components through a semantic role labeling technology, calculating statement similarity of to-be-fused content by applying a semantic matching technology based on a cosine similarity algorithm and a word vector model, and optimizing a content logic structure so as to eliminate redundant information; and finally, the follow-up visit content of this time is formed. And based on a preset follow-up plan, the follow-up content of this time is combined with the features extracted from the health data of the patient, and a voice notification strategy is generated by adopting a reinforcement learning algorithm. The follow-up visit management method has the following effect that the accuracy and efficiency of follow-up visit management of the patient are improved.

The invention relates to hallucination detection and processing of domain-specific dialog systems based on large language models. Systems and methods for detecting and reducing illusion and improving reliability of LLM-based QA systems are described herein. In particular, the present disclosure provides a set of illusion detection and processing methods that address different types of illusions of a QA system. The illusion detection method described herein includes comparing natural language answers to contextual information by means of sentence similarity estimation and keyword matching. The illusion processing method described herein includes removing an illusion sentence from the natural language answer or regenerating the natural language answer using better contextual information depending on a level of illusion detected in the natural language answer. A hybrid framework is also provided that systematically combines a hallucination detection method and a hallucination processing method into one system to achieve optimal hallucination reduction performance for a QA system.

The invention provides a database operation and maintenance knowledge graph construction method and system for optimizing entity alignment, and the system comprises 1, a data preprocessing module which carries out the data cleaning of unstructured and semi-structured data; 2, a knowledge extraction module which extracts valuable information and knowledge from the cleaned data, including entities, relationships and attributes; the entity alignment module is used for aligning the same or similar entities in different data sources; and 4, a knowledge representation and storage module which performs standardized representation on the extracted and aligned knowledge and stores the knowledge in a knowledge base. According to the method, the context attention weight is introduced, and the semantic similarity between the words is calculated by calculating the attention weight of the context word vector of each word to the target word vector, so that the accuracy of similarity calculation is improved; and a hybrid editing distance algorithm is utilized to bring the semantic similarity among the words into an editing distance calculation process, so that the calculation method of the overall sentence similarity is optimized. Insertion, deletion and replacement operations of characters are considered, and word similarity calculation of a semantic level is also combined, so that sentence similarity calculation has a semantic understanding capability.

This invention discloses a long text matching method combining noise filtering and a divide-and-conquer strategy. The method includes: constructing a long text matching model, which comprises a keyword extraction layer, an association extraction layer, and a filtering layer. The keyword extraction layer extracts keywords from the text; the filtering layer filters noise from the text based on sentence similarity to obtain a denoised text sequence; and the association extraction layer further removes keywords from the denoised text sequence to obtain the remaining associated text. The long text matching model is trained with the optimization objective of minimizing a set overall loss function, which reflects the global matching distribution and combines the keyword and association matching distributions. For the target text, real-time text matching is performed using the trained long text matching model. This invention improves the generalization ability and accuracy of text matching.

The present invention relates to a real-time semantic similarity matching design method based on a cosine similarity algorithm. A Java Web application end synchronizes data from an application database MySQL to a Kafka message queue cluster in real time, and completes real-time extraction and conversion of incremental data from the application database MySQL; a Python Web consumer obtains data through client consumption, and uses a cosine similarity algorithm to retrieve the semantic similarity of texts. The advantages of the present invention are: the text vector generated by the sentence embedding method contains the semantic relationship between words, can effectively measure the semantic similarity between texts, solves the problem of repeated verification of test questions, designs and implements a test question entry and architecture of a test paper system based on text semantic similarity, and has low computational complexity; the sentence similarity adopts a cosine similarity algorithm, the dynamic capture of change data adopts a Debezium low-latency streaming processing platform, and the Kafka message queue cluster realizes the extraction and storage of the test question stem stream data information.

The present application discloses a method for calculating sentence similarity, a related device, and a computer storage medium, wherein the method includes: obtaining two sentences to be compared and a decay factor; vectorizing phrases in the two sentences to be compared respectively to obtain a word vector for each phrase in the two sentences to be compared; determining the word vector similarity between the two sentences to be compared based on the word vector of each phrase in the two sentences to be compared; generating a syntactic structure binary tree of each sentence to be compared according to a preset sentence structure division rule; and determining the sentence similarity between the two sentences to be compared based on the decay factor, the syntactic structure binary tree of each sentence to be compared, and the word vector similarity between the two sentences to be compared. Compared with the prior art method of calculating the similarity between sentences based only on the syntactic structure binary tree, the present application is more accurate.

Embodiments of the present application disclose a sentence processing method and related device, which are applied to an artificial intelligence scene or a vehicle-mounted scene. For a target sentence and a plurality of first candidate sentences, a sentence similarity between each first candidate sentence and the target sentence is calculated. When it is determined based on the sentence similarity that the plurality of first candidate sentences do not match the target sentence, the plurality of first candidate sentences are segmented to obtain a plurality of second candidate sentences. A plurality of first to-be-fused sentences are determined from the plurality of second candidate sentences according to an importance parameter value of each second candidate sentence for the target sentence. The plurality of first to-be-fused sentences are fused according to a dependency relationship between words in the plurality of first to-be-fused sentences to obtain a fused sentence matched with the target sentence. The method can meet the actual demand for obtaining a matched sentence of the target sentence even when the plurality of first candidate sentences do not match the target sentence, thereby improving the processing effect of application scenarios such as a question and answer system and a search system.

The present application discloses a method for calculating sentence similarity, a related device, and a computer storage medium, wherein the method includes: obtaining two sentences to be compared and a decay factor; vectorizing the phrases and negations in the two sentences to be compared respectively to obtain word vectors for each phrase and negation in the two sentences to be compared; determining the word vector similarity between the two sentences to be compared based on the word vectors for each phrase and negation in the two sentences to be compared; generating a dependency tree for each sentence to be compared according to a preset dependency syntactic rule; and determining the sentence similarity between the two sentences to be compared based on the decay factor, the dependency tree for each sentence to be compared, and the word vector similarity between the two sentences to be compared. Compared with the prior art method of calculating the similarity between sentences based only on the dependency tree, the embodiments of the present application are more accurate.

The disclosure herein describes a system and method for attentive sentence similarity scoring. A distilled sentence embedding (DSE) language model is trained by decoupling a transformer language model using knowledge distillation. The trained DSE language model calculates sentence embeddings for a plurality of candidate sentences for sentence similarity comparisons. An embedding component associated with the trained DSE language model generates a plurality of candidate sentence representations representing each candidate sentence in the plurality of candidate sentences which are stored for use in analyzing input sentences associated with queries or searches. A representation is created for the selected sentence. This selected sentence representation is used with the plurality of candidate sentence representations to create a similarity score for each candidate sentence-selected sentence pair. A retrieval component identifies a set of similar sentences from the plurality of candidate sentences responsive to the input query based on the set of similarity scores.

The embodiment of the present invention provides a text classification recommendation processing method and system, wherein the method includes: clustering all titles based on the semantic similarity between each title and other titles; segmenting the main text under the titles clustered into one category; determining the template title and the key retained segments in the target main text corresponding to the template title based on the sentence similarity between each segment and all segments under other titles; and combining the template title and the key retained segments into a recommended template and saving them to a template library. The embodiment of the present invention proposes a clustering method for text titles based on a new semantic similarity comparison method, and finally combines to form a recommended template, so that the model can learn more semantic information and make a deeper semantic comparison; based on the clustering text classification method, innovative universal rules are added to enable text classification to achieve better results.

The present inventive concept provides for a method for automated decision modelling from text including obtaining a text corpus including a policy. Terms and syntax are identified within the text corpus related to the policy. Sentence similarities and co-references based on the terms and syntax are identified. Discourse and sentence level semantic parsing is performed based on the terms and the sentence similarities and the co-references using machine learning. A decision model template is generated based on the discourse and semantic parsing, and the decision model template is transformed into an automated decision model.

The present application discloses a method for calculating sentence similarity, a related device, and a computer storage medium, wherein the method includes: obtaining two sentences to be compared and a decay factor; vectorizing the phrases and negations in the two sentences to be compared respectively to obtain a word vector for each phrase and negation in the two sentences to be compared; determining the word vector similarity between the two sentences to be compared based on the word vector of each phrase and negation in the two sentences to be compared; generating a syntactic structure binary tree of each sentence to be compared according to a preset sentence structure division rule; and determining the sentence similarity between the two sentences to be compared based on the decay factor, the syntactic structure binary tree of each sentence to be compared, and the word vector similarity between the two sentences to be compared. Compared with the prior art method of calculating the similarity between sentences based only on the syntactic structure binary tree, the present application is more accurate.

The invention discloses a special field keyword extraction method in combination with semantic information, and belongs to the field of corpus recognition processing. Comprises: an expert constructs a domain dictionary; automatically discovering new words from the domain literature through a new word discovery algorithm, and inputting the new words into the domain dictionary after manual confirmation so as to continuously update the domain dictionary; loading the domain dictionary to a word segmentation device; a keyword is calculated in the target corpus through a TextRank algorithm, and a weight value of the keyword is obtained; replacing a keyword part in an original sentence of the target corpus with MASK to form an MASK sentence; vectors of the MASK sentence and an original sentence are extracted through a BERT algorithm, and the similarity value of the MASK sentence and the original sentence is calculated through the extracted vectors; and carrying out weighted calculation on the weight calculated through the TexkRank in the step (4) and the sentence similarity value extracted in the step (6) to obtain a final weight value. According to the method, the keywords of the scene in the special field can be accurately extracted, and data support is provided for upper-layer applications such as document association analysis, retrieval and situation display.

Systems and methods are described herein for detecting and reducing hallucinations and improving the reliability of an LLM-based QA system. Particularly, the disclosure provides a set of hallucination detection and handling approaches that address different types of hallucinations of the QA system. The hallucination detection approaches described herein include comparing a natural language answer with context information by way of sentence similarity estimation and keyword matching. The hallucination handling approaches described herein include removing hallucinated sentences from the natural language answer or regenerating the natural language answer using better context information, depending on a level of hallucination detected in the natural language answer. A hybrid framework is also provided that systematically combines the hallucination detection approaches and hallucination handling approaches into one system to achieve an optimal hallucination-reduction performance for the QA system.

To support easy extraction of appropriate similar cases.SOLUTION: A similar case extraction support system 1 comprises: storage means for storing sentences representing the contents of items having correlations and content sentence vectors where the sentences have been converted to vectors; vector generation means for converting a sentence into a vector using a transformer-based model; calculation means for calculating a degree of similarity between an input sentence vector representing a vector of the input sentence converted by the vector generation means and a content sentence vector; order determination means for determining an order of the sentences representing the contents in accordance with the degree of similarity; and display means for displaying results including the sentences representing the contents, the degrees of similarity, and the orders.SELECTED DRAWING: Figure 1

The present application relates to a method, apparatus, computer device, and storage medium for abstract extraction. The method includes: obtaining a text feature matrix of a text keyword graph structure, a text sentence similarity graph structure, a text sentence order graph structure, and sentence position information from the text to be extracted, superimposing the text keyword graph structure, the text sentence similarity graph structure, and the text sentence order graph structure to obtain a heterogeneous graph structure of text features, and obtaining an abstract extraction result through the heterogeneous graph structure and the text feature matrix. By using this method, graph structures with different features can be obtained, a heterogeneous graph structure with multi-feature fusion can be constructed through the graph structures with different features, and further, the text feature matrix of sentence position information is fused to implement text abstract extraction. The specific process considers different text features, can be applied to texts with complex structures, and has good scalability, thereby improving the accuracy of abstract extraction.

The present application discloses a method for calculating sentence similarity, a related device and a computer storage medium, wherein the method includes: obtaining two sentences to be compared and a decay factor; vectorizing the phrases in the two sentences to be compared respectively to obtain the word vector of each phrase in the two sentences to be compared; determining the word vector similarity between the two sentences to be compared based on the word vector of each phrase in the two sentences to be compared; generating a dependency tree for each sentence to be compared according to a preset dependency syntactic rule; determining the sentence similarity between the two sentences to be compared based on the decay factor, the dependency tree of each sentence to be compared, and the word vector similarity between the two sentences to be compared. The dependency tree in the embodiment of the present application can directly process the relationship between words in a sentence, and the number of nodes is greatly reduced, which is convenient for directly marking parts of speech and has the advantage of being concise and clear.

A system and method for document summarization generates summarized articles and risk factor categorizations for display at a graphical user interface (GUI) dashboard. A transaction monitoring system includes an adverse media dashboard pipeline for processing risk factor alerts and generating document summarizations for display at a user device. Document summarization extracts several sentences from a source text and stacks the sentences to create a summary. The method creates a vector representation of each sentence using a machine learning word embedding model and generates a sentence similarity matrix by computing cosine similarity values. A sentence graph creation algorithm creates a graph corresponding to the sentence similarity matrix and calculates importance scores used in selecting sentences for the document summary. The GUI dashboard includes first, second, third and fourth dashboard regions for displaying alert report records, media records, and graphical user interface layouts of document summaries and risk factor visualizations.

The present invention relates to artificial intelligence, and provides a method, device, computer equipment and storage medium for determining a text topic. The method comprises: dividing a target text into a plurality of text sentences according to a preset sentence segmentation rule; inputting each of the text sentences into a preset Bert model to obtain a sentence vector set composed of vectors of each of the text sentences; determining the cosine similarity between each summary sentence and each text sentence according to the sentence vector set, and obtaining a summary sentence similarity set including the cosine similarity between each summary sentence and each text sentence; determining the sentence importance value of each summary sentence according to the summary sentence similarity set; and determining the text topic of the target text from a plurality of summary sentences according to the sentence importance value. The sentence vector obtained by the Bert model in the embodiment of the present invention combines the semantics of the sentence context, so the accuracy of topic extraction can be improved by determining the text topic using the sentence vector obtained by the Bert model.

Some embodiments of the present application provide a method, apparatus, storage medium, and electronic device for obtaining sentence similarity. The method includes: obtaining a character weight for each character in a first sentence, and obtaining a character weight for each character in a second sentence; obtaining a sentence vector for the first sentence to obtain a first vector, and obtaining a sentence vector for the second sentence to obtain a second vector, wherein the first vector and the second vector are both multidimensional vectors; using the character weight as a constraint, obtaining a similarity value between the first sentence and the second sentence based on the first vector and the second vector. Some embodiments of the present application can calculate the similarity of any two sentences, quickly and accurately obtaining the similarity value of the two sentences.

The present invention provides a Chinese-Vietnamese neural machine translation method based on sample loss weighted denoising. By introducing confidence to quantify whether the current model has confidence in the translation training samples, the model confidence is used to determine the learning focus. Specifically, the confidence of each training example is measured by the model, and then the confidence score is used as a factor to measure the loss of its corresponding example. However, in the early stage of model training, the model confidence score is unreliable. The present invention uses the evaluation scores obtained by cross-language sentence similarity, sentence alignment, and sentence fluency as weights to constrain the model training loss. Through this method, the training data can be used more effectively, allowing the model to pay more attention to high-confidence training samples, thereby reducing noise interference and improving the performance of Chinese-Vietnamese neural machine translation.

The application provides an event graph construction reasoning method and device based on sentence similarity. The method comprises the following steps: extracting a cause-effect relationship and event description from an original text, and constructing an initial event graph; calculating the similarity between nodes, i.e. event descriptions, in the initial event graph, merging events with a cause-effect relationship and a similarity greater than a set threshold into an event cluster to obtain an event cluster graph; inputting an event description into the event cluster graph, determining an event cluster in which an event with the greatest similarity to the event description is located, returning a partial event cluster graph having a direct cause-effect relationship with the event cluster, and calculating a related cause-effect event probability. The application merges events into event clusters based on similarity calculation, replaces events as nodes in a network structure, simplifies the network structure, and improves the efficiency of finding causes and results.

The application discloses a kind of low-resource language machine translation method and device based on multilingual semantic understanding driving, method includes the following steps: constructing data set and pre-processing data set;Build semantic representation model, use the paraphrase sentence in paraphrase corpus to train semantic representation model, the vectorization representation of target language monolingual corpus is carried out, and semantic retrieval library is established;Mask language modeling target and inter-sentence similarity constraint are introduced, and enhanced semantic retrieval model is obtained;With enhanced semantic retrieval model, retrieve the candidate sentence similar to target language sentence semantics in semantic retrieval library, and carry out sorting screening, retain Top-K candidate sentence;After screening, candidate sentence is respectively paired with source language sentence, and enhanced parallel corpus is constructed and merged with original parallel corpus, to form extended training set;Neural machine translation model is trained using extended training set, and output translation text.The method of the application can improve semantic consistency, translation quality and cross-language generalization ability.

The invention discloses a text-based video retrieval method and device, equipment and a storage medium, and the method comprises the steps: obtaining text content used for retrieving videos, constructing sentence features and word features of the text content, carrying out video coarse screening based on the text content, obtaining a plurality of candidate videos, constructing video features and frame features of each candidate video, and carrying out video retrieval based on the video features and the frame features. And calculating video sentence similarity, video word similarity, frame sentence similarity and frame word similarity of the candidate videos, thereby calculating semantic similarity scores of the candidate videos, and determining the candidate video with the highest semantic similarity score as a retrieval output video of the text content. Therefore, by calculating the global and local multi-granularity similarity between the text and the matched video, the semantic similarity score finally used for evaluating the similarity with the text content is more accurate and reliable, so that the accuracy of video retrieval based on the text is improved.

This disclosure describes systems and procedures for detecting and reducing hallucinations and improving the reliability of an LLM-based QA system. Specifically, it provides a set of hallucination detection and handling approaches that address various types of hallucinations encountered by the QA system. The hallucination detection approaches described here involve comparing a natural language response with contextual information through sentence similarity estimation and keyword matching. The hallucination handling approaches described here involve removing hallucinated sentences from the natural language response or regenerating the natural language response using improved contextual information, depending on the hallucination level detected in the natural language response.A hybrid framework is also provided that systematically combines hallucination detection approaches and hallucination management approaches into a single system to achieve optimal hallucination reduction performance for the QA system.