Exobrain is the national representative R&D project for artificial intelligence (AI), with an aim of developing AI technology in the language-processing field so as to support humans with intelligent labor by ensuring the core intellectual property right (IPR) through development and standardization of the original technologies of ① natural language analysis, ② knowledge learning and accumulation, and ③question answering in natural languages, etc., to allow knowledge communication with experts.
Exobrain has been under development for 10 years, in 3 steps, with a focus on development of language intelligence SW capable of understanding a language, knowledge learning, and provision of services based on expert level knowledge. The primary research and development includes ① development of technologies for deep understanding of natural language, and processing the Korean language (morphemic analysis, named entity recognition, homonym analysis, dependency parsing, semantic role labeling, ellipsis resolution, reference resolution, etc.), ② development of knowledge learning and accumulation technologies; development of technologies of learning and accumulating knowledge in expert fields based on text big data, and ③ development of deep question answering technologies; development of technologies for deep question answering from which a correct descriptive answer can be inferred in fields of expertise such as law, patents, finance etc.
The Exobrain project is implemented by a consortium formed by major domestic research institutes for the purpose of developing a language intelligence SW capable of learning and accumulating languages and knowledge from text big data and providing knowledge in expert fields through communication with users. The detailed tasks of the consortium include the comprehensive project (hosted by ETRI), detailed task 1 (hosted by ETRI), detailed task 2 (hosted by Saltlux), and detailed task 3 (hosted by KAIST). The roles of each detailed task are as follows.
Exobrain aims to contribute to the creation of new markets in various industries by supporting humans with knowledge labor such as investigation and analysis of vast amounts of text used by a large number of humans and developing an expert system for quick and efficient decision making. This expert system could be used as an AI lawyer, an AI patent attorney, an AI consultant, etc.