Natural Language Processing means making computers process and understand human language. It refers to processing natural language through methods such as morphological analysis and semantic analysis so that computers can understand it, and then creating the outcome in forms that humans can understand such as text, voice, and image. Natural Language Processing can be largely divided into Natural Language Understanding and Natural Language Generation. Natural Language Understanding is the technology related to machine’s ability to understand natural language sentences such as machine reading, emotion analysis, semantic similarity measurement, and so on. On the other hand, Natural Language Generation refers to the technology of transforming the result of machine processing into the form of natural language. This includes conversation generation, story generation, and summary. The Language Lab studies Artificial Intelligence (AI) technology based on Deep Learning that enables computers to understand and speak human language. The goal is to develop document understanding and interactive AI solutions based on a natural language understanding model.
Language ModelThe model calculates the probability of consecutive words by modeling language phenomena. We develop our own models such as transformer-based BERT, T4, and ELECTRA and apply them to NLP applications. In particular, R&D is conducted on the super-giant language model EXAONE, which as recently been a hot issue for language models.
MRC/QAIt is a technology that enables machine to read documents and answer questions like humans do. We are expanding the research to the area of Open QA where machine can answer open questions in addition to questions related to a selected document.
Text AnalyticsOur plan is to secure the technology that can find and analyze meaningful information from unstructured text, and technologies for chatbots, such as information extract and document summarization as well as analysis of the intention behind questions.
Deep Document UnderstandingThe latest technologies such as properties information, test conditions, etc. from theses and patent information are analyzed, and specialized domain document understanding
technologies that generate databases are being procured.
Dialog SystemWe are studying the technology that enables AI to engage in human-level conversations. Our plan is to secure technologies that enable AI to understand the context of conversation, remember conversations and make inferences, and make action plans for each situation
Natural Language GenerationResearch is being carried out on technologies that the latest End-to-End Speech-to-Tect technologies which recognize and analyze the user’s voice in real-time and synthesize natural speech.
Meet our leader!
Stanley Jungkyu Choi, Head of Multimodal AI Research Group&Language Lab
“The Language Lab studies Machine Reading Comprehension, Text Analytics, Dialog System, Speech-to-Text, and Gigantic Language Models beyond GPT-3 with the vision of realizing Learning-by-Reading AI technology in which AI learns and infers documents by itself.
Among various research topics, we provide an environment where members can select the research they want as much as possible and conduct research freely. We are looking for passionate NLP researchers who will work with us in a wonderful research environment!”