Date: Sep 13, 1994
Venue: Hong Kong University of Science and Technology, Hong Kong
Speaker: Jeffrey D. Ullman
Title: Future Directions in Computer Science Research

1. Understanding and Generating Natural Language

   Grand Challenges:

   • Machine translation with high accuracy, in a significant domain.
   • Retrieval of information from text, answer question about content of document.

   Research Topics:

   • NLP - representing and modelling of information
   • Speech recognition

2. Computational Modelling and Prototyping

   Grand Challenges:

   • Concurrent engineering: system including design aids, design information resources, virtual assemble, virtual testing, device simulation. "Enginerring's associate": logical reasoning tools for guiding design.
   • Modelling of design of drugs or other chemicals
   • Weather prediction: a chaotic problem.

   Research Topics:

   • Numerical, non-numerical, hybrid computing to support modelling and prototyping.
   • Graphics and visualization - development of algorithm rather than waiting for super-computer.

3. Mobil, Autonomous, Cooperating robots

   Grand Challenges:

   • Robots in construction
   • Service robot for routine task such as delivery within buildings, clean-up, etc.
   • Robots for operating in hostile environment

   Research Topics:

   • Vision and sensing
   • Planning and programming
   • Mechanical technology
   • Adaptive computing

4. Making Information Intra-structure More Usable

   Grand Challenges:

   • New models of education: education product that is inexpensive, on-demand, interactive, with feedback from a live person in case of need.
   • Electronic library and information tools
   • Clinical decision and management systems.

5. Building Large Software System

   Grand Challenges:

   • Significant improvement in the efficiency and reliability of software of software development.
   • Use complex specification
   • Significant improvement in the maintainability of software.

   Research Topics:

   • Software tools and languages
   • Formal methods for software specification and verification.