Abstract:
A real-time utterance verification system according to the present invention includes a speech recognition unit configured to recognize an utterance of an utterer, a memory configured to store a program for verifying the utterance of the utterer in real time, and a processor configured to execute the program stored in the memory, wherein, upon executing the program, the processor generates and stores a list of the utterance of the utterer, performs a semantic analysis on each utterance included in the list, and generates, when the utterance is determined to be an inappropriate utterance for a listener as a result of the semantic analysis, utterance restricting information corresponding to the inappropriate utterance.
Abstract:
A method for establishing paraphrasing data for a machine translation system includes selecting a paraphrasing target sentence through application of an object language model to a translated sentence that is obtained by machine-translating a source language sentence, extracting paraphrasing candidates that can be paraphrased with the paraphrasing target sentence from a source language corpus DB, performing machine translation with respect to the paraphrasing candidates, selecting a final paraphrasing candidate by applying the object language model to the result of the machine translation with respect to the paraphrasing candidates, and confirming the paraphrasing target sentence and the final paraphrasing candidate as paraphrasing lexical patterns using a bilingual corpus and storing the paraphrasing lexical patterns in a paraphrasing DB. According to the present invention, the consistent paraphrasing data can be established since the paraphrasing data is automatically established.
Abstract:
Provided is a method of constructing a neural network translation model. The method includes generating a first neural network translation model learning a feature of source domain data used in an unspecific field, generating a second neural network translation model learning a feature of target domain data used in a specific field, generating a third neural network translation model learning a common feature of the source domain data and the target domain data; and generating a combiner combining translation results of the first to third neural network translation models.
Abstract:
An incremental self-learning based dialogue apparatus for dialogue knowledge includes a dialogue processing unit configured to determine a intention of a user utterance by using a knowledge base and perform processing or a response suitable for the user intention, a dialogue establishment unit configured to automatically learn a user intention stored in a intention annotated learning corpus, store information about the learned user intention in the knowledge base, and edit and manage the knowledge base and the intention annotated learning corpus, and a self-knowledge augmentation unit configured to store a log of a dialogue performed by the dialogue processing unit, detect and classify an error in the stored dialogue log, automatically tag a user intention for the detected and classified error, and store the tagged user intention in the intention annotated learning corpus.
Abstract:
Provided is a method of generating a language model using crossmodal information. The method includes: receiving language-based first modality information and non-language-based second modality information; converting the first modality information into a first byte sequence; converting the second modality information into a second byte sequence; converting the first and second byte sequences into a first embedding vector and a second embedding vector by applying an embedding technique for each modality; generating semantic association information between first and second modality information by inputting the first and second embedding vectors to a crossmodal transformer; and learning the language model by setting the generated semantic association information as training data.
Abstract:
Provided are a method and apparatus for constructing a compact translation model that may be installed on a terminal on the basis of a pre-built reference model, in which a pre-built reference model is miniaturized through a parameter imitation learning and is efficiently compressed through a tree search structure imitation learning without degrading the translation performance. The compact translation model provides translation accuracy and speed in a terminal environment that is limited in network, memory, and computation performance.
Abstract:
The present invention relates to a device of simultaneous interpretation based on real-time extraction of an interpretation unit, the device including a voice recognition module configured to recognize voice units as sentence units or translation units from vocalized speech that is input in real time, a real-time interpretation unit extraction module configured to form one or more of the voice units into an interpretation unit, and a real-time interpretation module configured to perform an interpretation task for each interpretation unit formed by the real-time interpretation unit extraction module.
Abstract:
Provided is a method of constructing a neural network translation model. The method includes generating a first neural network translation model learning a feature of source domain data used in an unspecific field, generating a second neural network translation model learning a feature of target domain data used in a specific field, generating a third neural network translation model learning a common feature of the source domain data and the target domain data; and generating a combiner combining translation results of the first to third neural network translation models.
Abstract:
The present invention relates to a spoken dialog system and method based on dual dialog management using a hierarchical dialog task library that may increase reutilization of dialog knowledge by constructing and packaging the dialog knowledge based on a task unit having a hierarchical structure, and may construct and process the dialog knowledge using a dialog plan scheme about relationship therebetween by classifying the dialog knowledge based on a task unit to make design of a dialog service convenient, which is different from an existing spoken dialog system in which it is difficult to reuse dialog knowledge since a large amount of construction costs and time is required.