Human-Social Information Sciences

Language and Information Science C01

  • Assoc. Prof. Itsumi Saito      
KeywordsNatural Language Processing, Deep Learning, Artificial Intelligence

Natural Language Processing for Enhancing Human-AI Collaboration

Our research aims to help AI work in collaboration with humans on various tasks by enabling computers to understand and process the natural language and related information that humans use daily, at a high level of knowledge processing.In order for AI to collaborate with humans at a high level, it needs to accurately understand and process the various information utilized in communication by humans, including natural language. Additionally, AI needs the ability to generate or edit text based on the processed content.


We are currently developing technology that can simultaneously understand not only natural language but also image and speech inputs, and that also has the ability to flexibly summarize or edit that information in accordance with instructions by humans.
Recent advancements in technology utilizing deep learning have significantly improved these capabilities but there are still challenges regarding the reliability of the generated information and the robustness against diverse inputs.
Furthermore, in order to collaborate efficiently with humans, it is important to have technologies based on understanding human cognition and communication, such as accurately understanding the intentions behind instructions and presenting information in a format that humans can easily comprehend. We aim to address these challenges and build AI systems that provide advanced support for human activities.
 
  • Information extraction and summarization for presentation.

  • Example of research process.