207 results about "Language understanding" patented technology

A method for assisting a user with one or more desired tasks is disclosed. For example, an executable, generic language understanding module and an executable, generic task reasoning module are provided for execution in the computer processing system. A set of run-time specifications is provided to the generic language understanding module and the generic task reasoning module, comprising one or more models specific to a domain. A language input is then received from a user, an intention of the user is determined with respect to one or more desired tasks, and the user is assisted with the one or more desired tasks, in accordance with the intention of the user.

Methods and systems for dynamic natural language understanding. A hierarchical structure of semantic categories is exploited to assist in the natural language understanding. Optionally, the natural language to be understood includes a request.

The invention relates to a method and an apparatus for separating processing for language understanding from an application and its functionality. The application for example contains functionality within a provided domain, whereby the domain and its functionality belongs to a speech controlled device such as a car-radio, a car-AC, a CD-player or almost any program controlled device.

A method for adaptively training a subject, using auditory processing of phonemes within command sentences, to improve the subject's listening comprehension, grammatical parsing, and serial memory is provided. The method utilizes a number of training installments, each designed for testing a particular aspect of the subject's language skills, all tied together by a common story. More specifically, installments are provided that narrate a story, test the subject's listening comprehension to the narrated story, test the subject's ability to grammatically parse increasingly difficult sentence structures, and test the subject's ability to select and manipulate graphical objects in response to auditory commands. Speech processing is used for the narration, as well as for commands within each test to allow the subject to more easily distinguish between similar sounding phonemes. As the subject improves his/her ability to correctly respond to the tests, the amount of processing applied to the commands is reduced, ultimately to the level of normal speech.

The invention provides distributed cognitive technology for an intelligent emotional robot, which can be applied in the field of multi-channel human-computer interaction in service robots, household robots, and the like. In a human-computer interaction process, the multi-channel cognition for the environment and people is distributed so that the interaction is more harmonious and natural. The distributed cognitive technology comprises four parts, namely 1) a language comprehension module which endows a robot with an ability of understanding human language after the steps of word division, word gender labeling, key word acquisition, and the like; 2) a vision comprehension module which comprises related vision functions such as face detection, feature extraction, feature identification, human behavior comprehension, and the like; 3) an emotion cognition module which extracts related information in language, expression and touch, analyzes user emotion contained in the information, synthesizes a comparatively accurate emotion state, and makes the intelligent emotional robot cognize the current emotion of a user; and 4) a physical quantity cognition module which makes the robot understand the environment and self state as the basis of self adjustment.

A method of probabilistic error-tolerant natural language understanding. The process of language understanding is divided into a concept parse and a concept sequence comparison steps. The concept parse uses a parse driven by a concept grammar to construct a concept parse forest set by parsing results of speech recognition. The concept sequence comparison uses an error-tolerant interpreter to compare the hypothetical concept sequences included by the concept parse forest set and the exemplary concept sequences included in the database of the system. A most possible concept sequence is found and converted into a semantic framed that expresses the intention of the user. The whole process is led by a probability oriented scoring function. When error occurs in the speech recognition and a correct concept sequence cannot be formed, the position of the error is determined and the error is recovered according to the scoring function to reduce the negative effect.

This invention concerns a method and system for monitoring an automated dialog system for the automatic recognition of language understanding errors based on a user's input communications in a task classification system. The method may include determining whether the user's input communication can be understood in order to make a task classification decision. If the user's input communication cannot be understood and a task classification decision cannot be made, a probability of understanding the user's input communication may be determined. If the probability exceeds a first threshold, further dialog may be conducted with the user. Otherwise, the user may be directed to a human for assistance. In another possible embodiment, the method operates as above except that if the probability exceeds a second threshold, the second threshold being higher than the first, then further dialog may be conducted with the user using the current dialog strategy. However, if the probability falls between a first threshold and a second threshold, the dialog strategy may be adapted in order to improve the chances of conducting a successful dialog with the user. This process may be cumulative. In particular, the first dialog exchange may be stored in a database. Then, a second dialog exchange is conducted with the user. As a result, a second determination is made as to whether the user's input communication can be understood can be conducted based on the stored first exchange and the current second exchanges. This cumulative process may continue using a third and fourth exchange, if necessary.

A natural language understanding method begins with a radiological report text containing clinical findings. Errors in the text are corrected by analyzing character-level optical transformation costs weighted by a frequency analysis over a corpus corresponding to the report text. For each word within the report text, a word embedding is obtained, character-level embeddings are determined, and the word and character-level embeddings are concatenated to a neural network which generates a plurality of NER tagged spans for the report text. A set of linked relationships are calculated for the NER tagged spans by generating masked text sequences based on the report text and determined pairs of potentially linked NER spans. A dense adjacency matrix is calculated based on attention weights obtained from providing the one or more masked text sequences to a Transformer deep learning network, and graph convolutions are then performed over the calculated dense adjacency matrix.

Techniques for detecting data anomalies in a natural language understanding (NLU) system are provided. A number of categorized sentences, categorized into a number of categories, are obtained. Sentences within a given one of the categories are clustered into a number of sub clusters, and the sub clusters are analyzed to identify data anomalies. The clustering can be based on surface forms of the sentences. The anomalies can be, for example, ambiguities or inconsistencies. The clustering can be performed, for example, with a K-means clustering algorithm.

A speech understanding apparatus includes a speech recognition unit which performs speech recognition of an utterance using multiple language models, and outputs multiple speech recognition results obtained by the speech recognition, a language understanding unit which uses multiple language understanding models to perform language understanding for each of the multiple speech recognition results output from the speech recognition unit, and outputs multiple speech understanding results obtained from the language understanding, and an integrating unit which calculates, based on values representing features of the speech understanding results, utterance batch confidences that numerically express accuracy of the speech understanding results for each of the multiple speech understanding results output from the language understanding unit, and selects one of the speech understanding results with a highest utterance batch confidence among the calculated utterance batch confidences.

A multi-lingual translation system that provides multiple output sentences for a given word or phrase. Each output sentence for a given word or phrase reflects, for example, a different emotional emphasis, dialect, accents, loudness or rates of speech. A given output sentence could be selected automatically, or manually as desired, to create a desired effect. For example, the same output sentence for a given word or phrase can be recorded three times, to selectively reflect excitement, sadness or fear. The multi-lingual translation system includes a phrase-spotting mechanism, a translation mechanism, a speech output mechanism and optionally, a language understanding mechanism or an event measuring mechanism or both. The phrase-spotting mechanism identifies a spoken phrase from a restricted domain of phrases. The language understanding mechanism, if present, maps the identified phrase onto a small set of formal phrases. The translation mechanism maps the formal phrase onto a well-formed phrase in one or more target languages. The speech output mechanism produces high-quality output speech. The speech output may be time synchronized to the spoken phrase using the output of the event measuring mechanism.

This language understanding apparatus includes a storage section, a phrase interpreter, and a bidding section. The bidding section reads information on a slot of a semantic frame corresponding to a concept indicated by an information read by the phrase interpreter and information on a concept corresponding to the slot by referring to the storage section, specifies, for each semantic frame read, from the concept representation detected by the phrase interpreter, a concept representation corresponding to a concept corresponding to each slot of the semantic frame or a subordinate concept of that concept, and a concept representation whose slot of another semantic frame is not bidden, and repeats bidding the slot for the specified concept representation, thereby generating slot allocation table data indicating a bidding result.

A language understanding device includes: a language understanding model storing unit configured to store word transition data including pre-transition states, input words, predefined outputs corresponding to the input words, word weight information, and post-transition states, and concept weighting data including concepts obtained from language understanding results for at least one word, and concept weight information corresponding to the concepts; a finite state transducer processing unit configured to output understanding result candidates including the predefined outputs, to accumulate word weights so as to obtain a cumulative word weight, and to sequentially perform state transition operations; a concept weighting processing unit configured to accumulate concept weights so as to obtain a cumulative concept weight; and an understanding result determination unit configured to determine an understanding result from the understanding result candidates by referring to the cumulative word weight and the cumulative concept weight.

Disclosed herein is a system, method and computer readable medium storing instructions related to semantic and syntactic information in a language understanding system. The method embodiment of the invention is a method for classifying utterances during a natural language dialog between a human and a computing device. The method comprises receiving a user utterance; generating a semantic and syntactic graph associated with the received utterance, extracting all n-grams as features from the generated semantic and syntactic graph and classifying the utterance. Classifying the utterance may be performed any number of ways such as using the extracted n-grams, a syntactic and semantic graphs or writing rules.

The invention provides a method and device for constructing a legal text syntax tree based on a deep neutral network. The method comprises the following steps: acquiring a training text of the legal service; performing word vector sparse representation on the training text; performing semantic annotation on the training text through the DNN according to the word vector sparse representation; performing dependency description on the training text according to the semantic annotation to construct the legal text syntax tree. The invention provides a systematic method to construct the syntax tree about technical language understanding in the legal service field, and the global-local training is used for solving the semantic bias problem in the legal language understanding.

The invention discloses a training method, a training device, a dialogue method and a dialogue system of a dialogue model. The training method comprises the steps of total error constructing, whereina total error function comprising a first error of a natural language understanding model and a second error of a strategy generating model are constructed simultaneously; joint training, wherein withthe target of reducing the total error function, a corpus sample is utilized to jointly train the natural language understanding model and the strategy generating model, input of the natural languageunderstanding model is dialogue sentences, output of the natural language understanding model is internal representation obtained by analyzing the dialogue sentences, input of the strategy generatingmodel at least comprises the output of the natural language understanding model, and output of the strategy generating model is motions aiming at the dialogue sentence. Compared with the prior art, when an error occurs on the natural language understanding model or the strategy generating model, the system can normally conduct dialogues, and the error transferring problem in a traditional methodcaused when the natural language understanding model and a dialogue management model are modeled respectively is solved.

A method of speech recognition and language-understanding analysis is provided. According to a segmental word-concept-tag compound N-gram model, an input speech is divided into a plurality of segmental phrases. Each segmental phrase is attached a tag to indicate whether said segmental phrase is a meaningful segmental phrase or a meaningless segmental phrase. The meaningless segmental phrases are deleted, and only the meaningful segmental phrases are reserved. The language-understanding analysis is carried out to the meaningful segmental phrases according to segmental sub-grammars.

The invention relates to a system and method for processing complex an SMS message of a mobile customer hotline SMS messaging service hall. The system comprises a login module for entering an operator reply module and a knowledge base maintenance module after passing validation of login information, an SMS message access module for receiving SMS messages incapable of being understood by the SMS messaging service hall, a natural language understanding module for understanding the content of an SMS message, a knowledge base module for managing mobile knowledge, an automatic replay module for automatically replaying SMS messages with high recognition rate, an operator replay module for manually reviewing and replying natural language understanding results by an operator, with the replaying content coming from the knowledge base, an answer knowledge search module for acquiring answer knowledge according to the SMS information and the natural language understanding results, a knowledge base searching module for inputting a natural language problem and searching the desired answer knowledge, a knowledge template instantiation module for obtaining a concrete knowledge answer, and a trigger operation module for being triggered to complete execution when the SMS content contains the execution requirement of a business action.

The invention belongs to the technical field of man-machine conversation, particularly relates to a multi-round spoken language understanding method, system and device based on conversation logic, andaims to solve the problem that an existing multi-round spoken language understanding method is low in historical conversation data utilization rate. The method of the system comprises the steps of obtaining current dialogue data and historical dialogue data; respectively encoding the current dialogue data and the historical dialogue data into an input vector and a memory vector through a bidirectional gated recurrent neural network; generating a context knowledge vector through a memory retrieval method based on an attention mechanism according to the input vector and the memory vector; and obtaining intention classification information and semantic slot filling information of the current dialogue data through a multi-round spoken language understanding model based on the context knowledge vector and the current dialogue data. Historical dialogue data can be efficiently utilized in multiple rounds of spoken language understanding, and the performance of language understanding in multiple rounds of scenes is improved.

The invention discloses a monitoring video multi-granularity marking method based on generalized multi-marking learning. The monitoring video multi-granularity marking method of the invention takes the backdrop of public security video monitoring content analysis and carries out a research according to video characteristic multi-layer acquisition and multi-granularity representation theory and method. The monitoring video multi-granularity marking method comprises steps of analyzing and extracting characteristics of different layers of an object in a video on the basis of a multi-marking learning theory and a deep learning theory , constructing a generalized multi-mark classification algorithm on the basis of a multi-mark learning theory and a deep learning theory, and characterizing a multi-granularity representation model of video information on the basis of a granular computing theory and a nature language understanding technology. The monitoring video multi-granularity marking method, targeting the monitoring video content field, carries out a research going deep into the system, constructs a multi-mark learning algorithm through the deep learning theory and can provide an effective theory and method to multi-layer video information. Through simulating the way that human recognize and describe the image, the monitoring video multi-granularity marking method establishes the multi-granularity video representation theory and method, provides a new idea to the video content analysis, and lays theory and application foundations for pushing development of future video monitoring intelligentalization.

A system and method for providing users with visual, English language comprehension, cognitive, or other limitations an intuitive audio interface and navigation method for use in automated systems which generates an official paper record of the session. A particularly useful application of the system is for voting systems whereby users listen to instructions and ballot choices via a telephonic device and use a standard 12-key telephone touchpad to enter selections. The selections made are printed in English and the voter's selection language, and also encapsulated in a two-dimensional identification symbol that is printed on an official paper ballot to be tabulated. Machine, ballot, and anti-fraud information is likewise encoded in the identification symbol for ensuring election integrity and complete accuracy. The barcoded paper ballots are the official ballots that serve as a physical audit trail.

The invention discloses a hierarchical semantic tree construction method and system for language understanding. The method mainly comprises the steps as follows: segmenting terms of a statement and loading a semantic knowledge base; recognizing all nodes of the statement according to an LV rule, and recognizing the level of the nodes according to semantic knowledge and term positions and collocations; generating a special node by punctuation at the end of the statement, and taking the special node as a root node of a semantic tree; merging the nodes according to generated node information, recognizing semantic side chunks of the statement, and taking a level-0 semantic side as a child node to be hung on the root node; circularly traversing all child nodes of the statement till no low-level semantic side exists, and taking the child nodes as leaf nodes to be hung on the child node. According to the hierarchical semantic tree construction method and system, under the condition that no syntactic resource exists, the semantic structure tree is obtained through semantic information and the term positions and collocations only, so that a computer can enter a deep semantic layer of a natural language, various processing of the natural language can be finished on the basis of understanding, the first step of semantic understanding of the natural language is realized, and the hierarchical semantic tree construction method and system can be applied to information retrieval, automatic abstraction, machine translation, text categorization, information filtration and the like.