Research

Research Contents

Development of technology that integrates deep learning knowledge
with knowledge base and probabilistic knowledge

Objective

The research designs a neural network model that integrates the symbolic knowledge of KBH built by humans and the symbolic knowledge of KBN built by neural networks. In addition, the research aims to design a model that provides high-level inference results by efficiently managing integrated knowledge through various conditions in Artificial Neural Hippocampus.

  • Konkuk university

    Technology integrating knowledge from existing knowledge bases with symbolic knowledge inferred from neural network models.

  • Sungkyunkwan university, Sogang university

    Technology to train untrained knowledge from deep learning with external symbolic knowledge

  • Wisenut, KAIST

    Expression and storage technology capable of efficient search and storage, merging, compression and oblivion of new entires

  • Korea university

    Probabilistic reasoning technique that integrates knowledge of neural network model extraction including uncertainty and symbolic knowledge