An entity recognition method and apparatus, an electronic device, a storage medium, and a computer program product are provided. The method includes: recognizing a to-be-recognized image to determine a preliminary recognition result for entities in the to-be-recognized image; determining, in response to determining that the preliminary recognition result includes a plurality of entities of a same category, image features of the to-be-recognized image and textual features of the plurality of entities; determining whether the plurality of entities is a consecutive complete entity based on the image features and the textual features, to obtain a complete-entity determining result; and obtaining a final recognition result based on the preliminary recognition result and the complete-entity determining result.

According to an embodiment, a recognition device includes a candidate detection unit, a recognition unit, a matching unit, and a prohibition processing unit. The candidate detection unit detects, from an input image, character candidates each being a set of pixels estimated to include a character. The recognition unit recognizes each of the character candidates and generates one or more recognition candidates each being a character of a candidate as a recognition result. The matching unit matches each of the one or more recognition candidates with a knowledge dictionary in which a recognition target character string is modeled, and generates matching results obtained by matching a character string estimated to be included in the input image with the knowledge dictionary. The prohibition processing unit deletes, from the matching results, a matching result obtained by matching a character string including a prohibition target character string with the knowledge dictionary.

According to an embodiment, a recognition device includes a detector, a recognizer, and a matcher. The detector is configured to detect a character candidate from an input image. The recognizer is configured to generate recognition candidate from the character candidate. The matcher is configured to match the recognition candidate with a knowledge dictionary and contains modeled character strings to be recognized, and generate a matching result obtained by matching a character string presumed to be included in the input image with the dictionary. Any one of a real character code that represents a character and a virtual character code that specifies a command is assigned to an edge. The matcher gives, when shifting a state of the dictionary in accordance with an edge to which the virtual character code is assigned, a command specified by the virtual character code assigned to the edge to a command processor.

Techniques are disclosed for generating a sequence of symbols based on input data for a neuro-linguistic model. The model may be used by a behavior recognition system to analyze the input data. A mapper component of a neuro-linguistic module in the behavior recognition system receives one or more normalized vectors generated from the input data. The mapper component generates one or more clusters based on a statistical distribution of the normalized vectors. The mapper component evaluates statistics and identifies statistically relevant clusters. The mapper component assigns a distinct symbol to each of the identified clusters.

Certain aspects of the present disclosure provide techniques for recommending trip purposes to users of an application. Embodiments include receiving labeled travel data from the application running on a remote device including a plurality of trip purposes. Embodiments include building a topic model representing words associated with a plurality of topics. Embodiments include training a topic prediction model, using the plurality of topics and one or more features derived from each of the plurality of trip records, to output a topic based on an input trip record. Embodiments include training a purpose prediction model, using the topic model and the plurality of trip purposes, to output a trip purpose based on an input topic. The trip purpose may be recommended to a user via a user interface of the application running on the remote device.

A method for customizing an optical character recognition system is disclosed. The optical character recognition system includes a general-purpose decoder configured to convert character images, recognized in a digital image, into text based on a general-purpose text structure. An application-specific customization is received. The application-specific customization includes an application-specific text structure that differs from the general-purpose text structure. A customized model is generated based on the application-specific customization. An enhanced application-specific decoder is generated by modifying the general-purpose decoder to, during run-time execution of the optical character recognition system, leverage the customized model to convert character images demonstrating the application-specific text structure into text.

A device may receive image data that includes an image of a document and lexicon data identifying a lexicon, and may perform an extraction technique on the image data to identify at least one field in the document. The device may utilize form segmentation to automatically generate label data identifying labels for the image data, and may process the image data, the label data, and data identifying the at least one field, with a first model, to identify visual features. The device may process the image data and the visual features, with a second model, to identify sequences of characters, and may process the image data and the sequences of characters, with a third model, to identify strings of characters. The device may compare the lexicon data and the strings of characters to generate verified strings of characters that may be utilized to generate a digitized document.

Examples described herein generally relate to systems and methods for handwriting recognition. In an example, a computing device may receive input corresponding to a handwritten word and apply first recognition model to the input. The first recognition model may be configured to determine a first confidence level of a first portion of the input is greater than a second confidence level of a second portion of the input. The computing device may also apply a second recognition model to the input, wherein the second recognition model is different from the first recognition model and combine results of the first recognition model and the second recognition model to determine a list of candidate words. The computing device may also output one or more candidate words from the list of candidate words.

The present teaching relates to method, system, and programming for responding to an image related query. Information related to each of a plurality of images is received, wherein the information represents concepts co-existing in the image. Visual semantics for each of the plurality of images are created based on the information related thereto. Representations of scenes of the plurality of images are obtained via machine learning, based on the visual semantics of the plurality of images, wherein the representations capture concepts associated with the scenes.

A method, a system, and a computer program product for detecting fake news based on a multi-task learning model. In an embodiment, a multi-task learning model is used to perform joint training on authenticity detection and topic classification of news to be detected, and authenticity of the news to be detected and a topic of the news to be detected are returned simultaneously. Through the implementation of the embodiment of the present invention, the authenticity of the news and the topic of the news can be detected simultaneously, and the accuracy of fake news detection and topic classification is improved.