Electronics and Telecommunications Research Institute

Soojong Lim

Jeong Mu Heo

Chung Hee Lee

Myung Gil Jang

Yeo Chan Yoon

Hyo-Jung Oh

Miran Choi

Hyunki Kim

Changki Lee

Pum Mo Ryu

YiGyu Hwang

Ki Young Lee

Young Kil Kim

Sang Kyu Park

Young-Ae Seo

Jong Hun Shin

Chang Ki Lee

Yi Gyu Hwang

Ji Ae Shin

Jihee Ryu

Yoonjae Choi

Jin Seong

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.

Method and apparatus for generating language model using crossmodal information

A question type and domain identifying apparatus includes: a question type identifier for recognizing the number of words of a user's question to identify whether the user's question is a query for performing information searching or a question for performing a question and answer (Q&A); a question domain distributor for distributing one of plural preset domain specialized Q&A engines, as a Q&A engine of the user's question based on the recognized word number; and a Q&A engine block, including the domain specialized Q&A engines, for selectively performing information searching or a Q&A with respect to the user's question in response to the distribution of the question domain distributor.

Question type and domain identifying apparatus and method

A topic map based indexing apparatus analyzes community Q/A lists to acquire Q/A analysis information, removes redundant answers depending on the Q/A analysis information, removes insignificant answers based on the degree of reliability, ranks answer lists, and extracts the highest ranking answer as a best answer, to thereby store, in a community Q/A topic map, index information containing the community Q/A lists and the Q/A analysis information. A topic map based searching apparatus analyzes a user question to acquire question analysis information, searches similar questions from community Q/A lists belonging to a specific topic node of a pre-stored community Q/A topic map, ranks the searched similar questions depending on the question analysis information, removes redundant answers among answers to the ranked similar questions, ranks the answers, and extracts the highest ranking answer as a best answer.

Topic map based indexing and searching apparatus

An apparatus for verifying training data using machine learning includes: a training data separation unit for separating provided initial training data into N training data and N verification data, where N is a natural number; a machine learning unit for performing machine learning on the separated training data to generate a training model; an automatic tagging unit for automatically tagging an original text of the verification data using the generated training model to provide automatic tagging results; and an error determination unit for comparing the verification data to the automatic tagging results to determine error candidates of the training data.

Apparatus and method for verifying training data using machine learning

A method for stabilizing a knowledge graph includes: generating a knowledge graph in which same entities in a semantic relation list between entities provided as an input are represented as a single node based on names and types of the entities; computing, on the knowledge graph, semantic similarities between all potential entity pairs of same entity types by comparing, for each potential entity pair, a type of relation associated with an entity in the entity pair and an opponent entity to the entity; and selecting, based on the semantic similarities, a representative entity from each of semantically similar entity pairs on the knowledge graph and integrating an opponent entity to the representative entity into the representative entity. The method further includes computing relation weighted values between the entities by using a graph analysis and statistic information, and adding the weighted values to the knowledge graph.

Apparatus and method for knowledge graph stabilization

Growing community of inventors

Daejeon, South Korea

Soojong Lim