18 results about "Sentence generation" patented technology

The disclosure provides a reply generation model training method and device and a reply sentence generation method and device, relates to the fields of artificial intelligence such as natural language processing and deep learning, and comprises the following steps: obtaining a first training sample of a candidate reply generation model to be trained and a sample dialogue context, a sample fact and a reference fact label in the first training sample; identifying whether the sample fact and the reference fact label are matched and determining a model training task of a current round of the candidate reply generation model according to an identification result; obtaining a training loss of the candidate reply generation model under the model training task; adjusting parameters of the candidate reply generation model according to the training loss and returning the candidate reply generation model after the parameter adjustment to obtain a next sample dialogue context for continuing training, until a target reply generation model is obtained after training is completed. The possibility of machine error replies caused by the inconsistency between the reply information and the fact information generated by the model is reduced, and the accuracy of the reply information is improved.

A factuality verification apparatus includes an input module configured to receive a written document, a filtering module configured to filter an untargeted sentence, which does not correspond to a target for factuality verification, a selection module configured to respectively divide a plurality of individual facts of one sentence into different sentences to generate atomic sentences respectively corresponding to the plurality of individual facts and select a core atomic sentence corresponding to a sentence including the fact-related content associated with the subject among the atomic sentences, and a verification module configured to perform factuality verification on whether the core atomic sentence is true or not, through comparison of the core atomic sentence and an evidence document capable of checking whether content of the core atomic sentence is true or false.

In order to classify, stably with high accuracy, a document to be classified, a document classification apparatus (1) includes: a hypothetical sentence generation section (11) that generates, for a candidate classification as which a document is to be classified, a plurality of hypothetical sentences, which are sentences related to the candidate classification; an entailment evaluation section (12) that subjects each of the plurality of hypothetical sentences to a process for evaluating whether the document entails a corresponding hypothetical sentence; and an evaluation result aggregation section (13) that determines, on the basis of a result of evaluation of each of the plurality of hypothetical sentences by the entailment evaluation section (12), a classification as which the document is to be classified.

The application discloses a voice outbound call guide sentence generation method and device, equipment and storage medium, relates to the technical field of intelligent conversation, the whole process combines double large language model cooperative processing, accurately identifies user history outbound objection, and generates customized outbound call guide sentence adaptively combined with task requirements and high-quality corpus, effectively improves outbound communication efficiency and business adaptability. The method comprises the following steps: obtaining the historical outbound call dialogue record of a target user, the current outbound call task target and the current task node information; inputting the historical outbound call dialogue record and the prompt word template into a first large language model to output structured data containing objection category labels, original objection sentences and objection contexts; inputting the objection category labels, the objection contexts, the current outbound call task target, the current task node information and a pre-constructed reference dialogue corpus into a second large language model to generate a voice outbound call guide sentence.

This invention proposes a paraphrase generation method based on a controllable latent space diffusion model, comprising the following steps: Step 1, constructing a paraphrase model based on a latent space diffusion model, and training the paraphrase model based on an existing paraphrase text dataset; the existing paraphrase text dataset contains the original sentence and its paraphrase; Step 2, dividing the paraphrase source sentence into key information and non-key information, and constructing enhanced semantic fragment information; Step 3, constructing a controller; Step 4, training the controller; Step 5, combining the paraphrase model based on a latent space diffusion model trained in Step 1 and the controller trained in Step 4 for collaborative reasoning to generate paraphrased sentences, thus completing the paraphrase generation task based on a controllable latent space diffusion model.

A system includes a processor that acquires observation information comprising environmental information including structures, moving objects, and luminance states in a plurality of spatial regions, acquires interview information provided by regional stakeholders as opinion information, generates an instruction sentence for predicting regional environmental characteristics for a plurality of future periods based on the observation information and the interview information, inputs the instruction sentence generated by the instruction sentence generation to a generative AI model, and acquires prediction information indicating the regional environmental characteristics for the plurality of future periods from the generative AI model, and generates and outputs indicator information applicable to resource allocation policy or business planning based on the prediction information.

The application discloses a PromQL sentence generation method and device, equipment and a readable storage medium, the method comprises the following steps: identifying the query requirement of the user's question; recalling the target PromQL sentence and the sentence analysis text block matched with the query requirement from the vector database and the historical log library, and constructing the reference context; using the sentence generation model, generating the PromQL query sentence based on the reference context; executing the PromQL query sentence to obtain the execution result; updating the reference context when the PromQL query sentence cannot meet the query requirement; returning to the step of generating the PromQL query sentence for meeting the query requirement until the latest execution result meets the query requirement. It can be seen that the application can realize the conversion from natural language to PromQL sentence, effectively improve the PromQL generation efficiency, accuracy and usability.

The application provides a sentence generation method, a sentence generation device and an electronic device. The sentence generation method comprises the following steps: obtaining a first prefix tree corresponding to first input information, wherein the first prefix tree is used to limit a generation space corresponding to the first input information; generating at least one first reply sentence according to the first input information and the first prefix tree, wherein each character of the first reply sentence is located on a same branch of the first prefix tree; and outputting the at least one first reply sentence.

We provide the system. [Solution] An observation information acquisition means for acquiring observation information consisting of environmental information including structures, moving objects, and brightness states in multiple spatial regions, A means for acquiring interview information consisting of opinion information provided by local stakeholders, Instructions generation means for generating instructions for predicting regional environmental characteristics multiple periods in the future based on the observation information and the interview information, A prediction information acquisition means inputs the instruction sentences generated by the instruction sentence generation means into a generation AI model and obtains prediction information from the generation AI model that shows the regional environmental characteristics after the multiple periods, An indicator information output means generates and outputs indicator information applicable to the formulation of resource allocation policies or business plans based on the aforementioned forecast information. A system that includes this.

The application discloses an intent structured SQL sentence generation method based on a large language model, and comprises the following steps: obtaining a natural language query sentence input by a user, database schema information corresponding to a target data source, and semantic layer resources; performing normalization processing on the natural language query data; inputting the normalized query sentence into the large language model; constructing a retrieval unit set based on a query intent intermediate representation, and performing retrieval in a semantic layer resource set; based on the query intent intermediate representation and the semantic retrieval result set; generating an SQL main skeleton based on a compilable query intent intermediate representation and a template; calling an improved PICARD algorithm to perform a parser-in-the-loop multi-dimensional constraint incremental decoding processing on a to-be-generated SQL structure; and filling the constrained SQL segment into the to-be-generated SQL structure to generate a target SQL sentence. The application adopts an intent structuring and constraint decoding mechanism to realize controllable generation from natural language to SQL.

The application relates to a remote sensing image description generation method based on external knowledge retrieval enhancement, which comprises the following steps: step 1, visual feature coding; step 2, external knowledge retrieval enhancement; step 3, cross-modal generation decoder; and step 4, multi-task joint optimization. The application has the beneficial effects that: the retrieval enhancement generation paradigm is used; through manifold alignment based on principal component analysis and a diversity perception reordering strategy, high-confidence semantic knowledge is efficiently extracted and fused from an external knowledge base, and the limitation of a parameterized model in fine-grained semantic cognition is effectively made up. Meanwhile, a cross-modal attention entropy regularization mechanism is adopted, the cross-modal alignment quality in the generation stage is improved, the attention collapse phenomenon is inhibited, and the model is encouraged to balance attention to global information of an image. In addition, a joint loss function is used in combination with a focal loss and length normalization, and the long-tail vocabulary prediction and short sentence generation problems are systematically alleviated.

The application discloses a method and system for text excerpting by using voice, which can make a user quickly and accurately excerpt a corresponding book fragment by reading text content and with the aid of automatic completion technology. The technical scheme is as follows: converting information in a user voice signal into text; obtaining candidate sentences from preselected book contents according to the converted text; generating a candidate fragment according to the obtained candidate sentences; and generating a book fragment based on the candidate fragment.

The application discloses a natural language query retrieval system based on deep learning, comprising: a metadata acquisition module for acquiring metadata information of a target structured database; a field semantic generation module for generating a field-level semantic representation; a semantic card generation module for updating the field-level semantic representation and encapsulating the same as a semantic card; a query semantic generation module for generating a query-level semantic representation; a field selection module for generating a target field set; a constraint construction module for constructing a query constraint graph; a sentence generation and verification module for generating a structured query statement based on the query constraint graph and performing semantic consistency verification; and an execution and clarification module for generating an updated structured query statement, which realizes accurate mapping between a natural language query and structured data and improves the accuracy of field selection and constraint analysis in a complex natural language query scenario.

This disclosure presents embodiments of a response statement generation method, apparatus, electronic device, computer-readable medium, and program product. One specific implementation of the method includes: determining the user intent category to which the user input statement belongs; determining the response intent category based on the user intent category and historical dialogue information corresponding to the user input statement; selecting response materials matching the user intent category and the response intent category from a preset response material library; determining a prompt word template based on the response intent category and the selected response materials; and generating a response statement for the user input statement using the prompt word template. This implementation is related to artificial intelligence and helps improve the intelligence of dialogue with users.

Embodiments of the present specification provide an information query method, which comprises: obtaining a query request, wherein the query request contains a query text; inputting the query text into a query sentence generation model to obtain a target query sentence corresponding to the query text, wherein the query sentence generation model is obtained based on a sample set, the sample set contains sample texts and sample query sentences, the sample query sentences are obtained based on different sample expansion tasks, and the sample texts are obtained by inputting the sample query sentences into a text generation model; and searching for target information from a database based on the target query sentence. Since the sample set is obtained based on the sample expansion task and the text generation model, the query sentence generation model trained based on the sample set containing rich samples has higher accuracy in generating the target query sentence, and the target information obtained based on the target query sentence is more accurate.

The present application relates to the field of natural language processing, and particularly relates to a biomedical event trigger extraction method and system based on hypergraph neural network. The method of the present application first preprocesses an unstructured biomedical data set; then obtains feature embeddings of all text information in the biomedical corpus through a pre-trained model to obtain vector representations of each word; then generates a corresponding hypergraph structure for each sentence; then inputs the feature embeddings and the hypergraph structure of each sentence into a hypergraph convolutional neural network to define a cross-entropy loss function to train the model; finally, trigger word detection is performed on an unlabeled test set. The present application differs from existing methods in that a bidirectional LSTM is used to aggregate context information in each sentence, but instead uses a hypergraph structure to aggregate context information, which is highly effective and achieves the purpose of improving the accuracy of trigger word extraction in biomedical text.

The application discloses a kind of based on template sentence enhancement's paraphrase sentence generation method, steps are as follows: step 1: obtain original sentence, and obtain the target paraphrase sentence of original sentence in parallel corpus, use the method of rule to retrieve 1 sentence from parallel predata with the highest similarity with target paraphrase sentence as optimal sentence as reference;Step 2: different part-of-speech special symbols are used to replace noun, verb, adjective and adverb in optimal sentence to construct template sentence;Step 3: original sentence and template sentence are spliced through special symbol and joint, and the spliced sentence is used as the input of paraphrase generation model, and the paraphrase generation model outputs several paraphrase sentences;Step 4: the similarity of several paraphrase sentences and original sentence in grammatical structure and semantics is calculated, and the highest similarity paraphrase sentence is used as the optimal paraphrase sentence of original sentence.The application avoids the information dependence of model training phase in template sentence by masking the words of related part-of-speech in template sentence.