8 results about "Encyclopedia" patented technology

The application provides a kind of construction method of intelligent question answering system based on computer discipline knowledge graph, first constructs computer discipline ontology, then utilizes crawler technology to collect unstructured data such as electronic books and baidu encyclopedia, completes knowledge processing based on Colabeler tool, BiLSTM-CRF, BiLSTM-ATT, stores the extracted entity, relationship and other data to Neo4j, and finally establishes intelligent question answering system based on backend Flask framework and front-end React framework. First, the keywords of the question are extracted, and the template is matched. If no answer is matched, the relevance of the question is analyzed, and the question is stored in the relational database. The teacher answers, extracts the effective information of the teacher's answer using the knowledge extraction model, and supplements it to the knowledge graph to update the knowledge graph. The application has three characteristics of data structuring, storage refinement and intelligent question answering, which solves the problems of scattered and irregular computer discipline data, coarse discipline knowledge granularity and inaccurate search results.

The application provides a social user location prediction method based on a pre-trained language model and a knowledge graph, which comprises the following steps: obtaining a valid text set and a valid entity set published by a first social user; using a pre-trained language model to encode the valid text and the valid entity respectively to obtain semantic representations of the valid text and the valid entity; calculating an anonymous relationship representation between the valid entity and the corresponding valid text according to the semantic representations of the valid text and the valid entity; extracting a k-hop subgraph of the valid entity from a geographic encyclopedia knowledge graph; constructing a joint graph of the first social user according to the entity, the valid text, the valid entity, the anonymous relationship representation and the k-hop subgraph of the first social user; and predicting the location of the first social user by using a progressive relationship graph neural network according to the joint graph. The application can mine the potential correlation between the user and the location, has strong generalization, and can perform interpretable reasoning.

The invention relates to the technical field of artificial intelligence and electronic commerce, and particularly discloses an encyclopedia information analysis and product matching system based on questions and answers. The system comprises a question and answer interface module, an encyclopedia information analysis module, a product matching module and an output module. The question and answer interface module is responsible for receiving and analyzing natural language questions of a user; the encyclopedia information analysis module extracts key knowledge points from the multi-source encyclopedia data; the product matching module generates a product combination scheme based on knowledge point matching; and the output module presents the result in a structured form. According to the system, accurate knowledge point extraction is realized through semantic vector calculation, a multi-dimensional evaluation system is adopted to generate a matching scheme, and personalized services are provided in combination with user portraits. According to the method, the knowledge query and the product recommendation are organically combined, the information processing efficiency and the recommendation accuracy are improved, and stable operation in a high-concurrency scene is supported. The system can be widely applied to the fields of electronic commerce, intelligent customer service and the like, and has important practical value.

The present disclosure provides an information display method, device, computer apparatus and storage medium, wherein the method comprises: receiving an access request for book encyclopedia information of a target book; acquiring the book encyclopedia information of the target book, wherein, the book encyclopedia information comprises a plurality of information modules, each information module corresponding to at least one book attribute dimension, and the book encyclopedia information belonging to each book attribute dimension is determined according to user's original innovative information and / or information obtained by automatically identifying book-related content of the target book; acquiring and displaying a book encyclopedia page matching book category of the target book, and displaying the book encyclopedia information of each information module in the book encyclopedia page.

This invention relates to a time-weighted difference-aware response generation method and system based on auto-step learning. The method includes the following steps: Step A: Collect natural, open-domain dialogues from social media software, and use a retrieval system to search for relevant background knowledge of the dialogue in an encyclopedia, and label real responses and real knowledge to construct a training dataset. DE Step B: Using the training dataset DE Step C involves training a time-weighted difference-aware deep learning network model M based on a self-synchronizing learning strategy. This model selects necessary background knowledge and generates responses based on the selected background knowledge. Step C involves inputting the dialogue context and background knowledge base into the trained deep learning network model M to generate corresponding responses. This method and system help improve the accuracy of generated responses.

This invention provides an optimization method for a temporally dependent data prediction model based on agent trajectory feedback, applicable to tasks such as webpage evidence text collection, open encyclopedia question answering, enterprise knowledge base question answering, or retrieval enhancement generation. The method collects multi-round execution trajectories formed by the agent's interaction with the retrieval system during task execution. Initial supervised samples are constructed based on the candidate document set returned by the search action and subsequent browsing actions. Positive samples are filtered using post-browsing inference text and an inference-evidence consistency judgment model. Relevance strength weights are estimated based on inference length, evidence citation count, number of fact entries, and changes in subsequent actions. The retrieval model is trained using temporally gated coding and a weighted contrastive learning objective function. The optimized retrieval model is then deployed back to the agent system for continuous closed-loop updates, improving the execution efficiency of webpage evidence collection, evidence retrieval, and complex question answering tasks.

This invention discloses a method and apparatus for generating negative samples based on lexical and semantic similarity in the Chinese domain. The method includes: extracting entities and all relations within one hop of each entity from a Chinese encyclopedia knowledge graph; concatenating the relations to form descriptive text for the entities; using Word2Vec trained on Chinese corpora to perform vector representation of the entities and descriptive text; performing similarity matching between the target matching entity and all entities, returning entities that meet the criteria as a set of semantic similarity candidate entities; calculating the edit distance between the target matching entity and all entities again, returning entities that meet the criteria as a set of lexical similarity candidate entities; and selecting high-confusion negative samples based on the semantic similarity candidate entity set and the lexical similarity candidate entity set. This invention combines lexical similarity and semantic similarity for negative sample selection, which can accelerate the convergence speed of the model and enable the model to train a good decision boundary.

The present application belongs to the technical field of knowledge mining, and particularly relates to a knowledge mining method based on cross-model and cross-language knowledge modeling. In the implementation process of the mining method, multi-language news knowledge is mined, different knowledge models constructed by the method have certain isolation and certain correlation, for different source language materials and knowledge models, a combination of a supervised knowledge extraction method and an unsupervised knowledge extraction method is adopted for knowledge extraction, and the knowledge obtained in the knowledge extraction process is mapped with the knowledge models, at the cross-language knowledge fusion level, a cross-language knowledge base is constructed based on an open source encyclopedia and an open source knowledge base, and knowledge fusion alignment is performed based on cross-language linking. Finally, cross-model and cross-language knowledge modeling and knowledge mining are completed, and a required cross-language news knowledge base is obtained.