Recent advances in deep learning and knowledge acquisition have given rise to a body of new approaches for utilizing knowledge graph in different applications, such as information retrieval and question answering. Based on the joint research with the National Institute of Advanced Industrial Science and Technology (AIST), Yu laboratory (Information Intelligence Lab) is conducting “Technology Development Project on Next-Generation Artificial Intelligence” (FY2020-FY2024), which we focus on how to develop a large-scale and effective knowledge graph that can support open-ended commonsense reasoning.
In order to develop a large-scale and effective knowledge graph that can support open-ended commonsense reasoning (namely, the task of answering a commonsense question based on a corpus of raw texts), many fundamental research topics have to be comprehensively explored. For example, the possible research topics may include (but not limited to): (1) effective and efficient annotation of raw text, such as named entities and verbs; (2) accurate relation identification among entities; (3) effective and efficient embedding of knowledge graph; (4) Joint generation and optimization of knowledge graph; (5) effective integration and evaluation of knowledge graph across different applications.
For this project, the students who have the following desires are greatly welcomed to join us,

  • highly self-motivated students aiming for solving real problems based on machine learning or deep learning
  • having the capability of reading English papers

After joining our lab, at first, we will help him/her to identify a suitable research topic. Next, our lab will assign a series of basic studies, such as machine learning and programming based on PyTorch. At the same time, our lab will help the student to establish a reasonable research plan, which enables him/her to conduct research step by step. The supervisor will provide in-time and necessary helps during the periodic seminar and discussion.