An apparatus for generating an application document is provided. The apparatus retrieves a plurality of candidate questions from at least one database, each of the plurality of candidate questions corresponding to one of a plurality of entities, and each of the plurality of entities being different from each other, removes substantively similar candidate questions among the plurality of candidate questions from the different entities; and generates a graphical user interface by aggregating remaining candidate questions among the plurality of candidate questions after the removal of the substantively similar candidate questions as application questions.

The present disclosure provides an information extraction method and apparatus, an electronic device and a readable storage medium, and relates to the field of natural language processing technologies. The information extraction method includes: acquiring a to-be-extracted text; acquiring a sample set, the sample set including a plurality of sample texts and labels of sample characters in the plurality of sample texts; determining a prediction label of each character in the to-be-extracted text according to a semantic feature vector of each character in the to-be-extracted text and a semantic feature vector of each sample character in the sample set; and extracting, according to the prediction label of each character, a character meeting a preset requirement from the to-be-extracted text as an extraction result of the to-be-extracted text. The present disclosure can simplify steps of information extraction, reduce costs of information extraction and improve flexibility and accuracy of information extraction.

The present disclosure provides an information extraction method and apparatus, an electronic device and a readable storage medium, and relates to the field of natural language processing technologies. The information extraction method includes: acquiring a to-be-extracted text; acquiring a sample set, the sample set including a plurality of sample texts and labels of sample characters in the plurality of sample texts; determining a prediction label of each character in the to-be-extracted text according to a semantic feature vector of each character in the to-be-extracted text and a semantic feature vector of each sample character in the sample set; and extracting, according to the prediction label of each character, a character meeting a preset requirement from the to-be-extracted text as an extraction result of the to-be-extracted text. The present disclosure can simplify steps of information extraction, reduce costs of information extraction and improve flexibility and accuracy of information extraction.

The present disclosure relates to concurrent learning of a relationship estimation model and a phrase generation model. The relationship estimation model estimates a relationship between phrases. The phrase generation model generates a phrase that relates to an input phrase. The phrase generation model includes an encoder and a decoder. The encoder converts a phrase into a vector using a three-piece set as learning data. The decoder generates, based on the converted vector and a connection expression or a relationship label, a phrase having a relationship expressed by the connection expression or the relationship label for the phrase. The relationship estimation model generates a relationship score from the converted vector, which indicates each phrase included in a combination of the phrases, and a vector indicating the connection expression and the relationship label.

The present disclosure relates to concurrent learning of a relationship estimation model and a phrase generation model. The relationship estimation model estimates a relationship between phrases. The phrase generation model generates a phrase that relates to an input phrase. The phrase generation model includes an encoder and a decoder. The encoder converts a phrase into a vector using a three-piece set as learning data. The decoder generates, based on the converted vector and a connection expression or a relationship label, a phrase having a relationship expressed by the connection expression or the relationship label for the phrase. The relationship estimation model generates a relationship score from the converted vector, which indicates each phrase included in a combination of the phrases, and a vector indicating the connection expression and the relationship label.

A method and apparatus for processing risk-management feature factors based on user generated content (UGC), an electronic device and a storage medium are disclosed, which relates to the fields of artificial intelligence and cloud computing. An implementation includes generating a feature expression of the UGC based on the UGC; and extracting the risk-management feature factors of the UGC according to a pre-generated risk-management-feature-factor extracting model and the feature expression of the UGC. According to the technology of the present application, the risk-management feature factors of a corresponding user may be extracted based on the UGC without depending on privacy information of the user, such as personal basic attributes, or the like, such that subsequent related processing actions of risk management may be facilitated, an acquiring way and an acquiring mode of the risk-management feature factors may be enriched effectively, and richer information of the risk-management feature factors may be acquired.

A method and apparatus for processing risk-management feature factors based on user generated content (UGC), an electronic device and a storage medium are disclosed, which relates to the fields of artificial intelligence and cloud computing. An implementation includes generating a feature expression of the UGC based on the UGC; and extracting the risk-management feature factors of the UGC according to a pre-generated risk-management-feature-factor extracting model and the feature expression of the UGC. According to the technology of the present application, the risk-management feature factors of a corresponding user may be extracted based on the UGC without depending on privacy information of the user, such as personal basic attributes, or the like, such that subsequent related processing actions of risk management may be facilitated, an acquiring way and an acquiring mode of the risk-management feature factors may be enriched effectively, and richer information of the risk-management feature factors may be acquired.

The present invention relates to an AI-based voice ordering system and a method therefor and provides a voice ordering method and system, the voice ordering method comprising: a first step of an ordering smart terminal standing by for voice data reception; a second step of the ordering smart terminal analyzing whether an input signal has been received by an input unit corresponding to a microphone activation button; and a third step of, if the analysis result indicates that an input signal has not been received, returning to the first step and, conversely, if an input signal has been received, the ordering smart terminal receiving a voice signal from a microphone, converting the voice signal into voice data of a preset format, and then transmitting the converted voice data to a voice ordering server via a host terminal connected to a network, so that analysis of text data is performed.

The present invention relates to an AI-based voice ordering system and a method therefor and provides a voice ordering method and system, the voice ordering method comprising: a first step of an ordering smart terminal standing by for voice data reception; a second step of the ordering smart terminal analyzing whether an input signal has been received by an input unit corresponding to a microphone activation button; and a third step of, if the analysis result indicates that an input signal has not been received, returning to the first step and, conversely, if an input signal has been received, the ordering smart terminal receiving a voice signal from a microphone, converting the voice signal into voice data of a preset format, and then transmitting the converted voice data to a voice ordering server via a host terminal connected to a network, so that analysis of text data is performed.

The present disclosure is directed to machine learning model architectures which provide full attention capability in each attention head while maintaining low computation and memory complexity. Specifically, according to one aspect of the present disclosure, example attention models provided herein can treat the self-attention mechanism as a conditional expectation over embeddings at each location and approximate the conditional distribution with a structured factorization. Each location can attend to all other locations, either via direct attention, or through indirect attention to group representations, which are again conditional expectations of embeddings from corresponding local regions.