以外行身份与外国学者讨论人工智能

武夷山

2003年5月17日

Dear Prof. Angelica:

Thank you very much for forwarding my message to Prof. Brockman.

I’m not a specialist in AI myself but it is a field I’m concerned with. In 1987, I translated Herbert A. Simon’s The Sciences of the Artificial into Chinese, in which one chapter is about AI. Last year, I read John Casti’s The Cambridge Quintet, and wrote a review of it and published the review in a nationally-circulated Chinese newspaper.

As a whole, I’m an optimist about the future of AI. In that sense, I won’t agree with Jaron Lanier’s over-pessimist standing. However, as an outsider of AI field, I do believe that the appropriate direction for AI research is to understand human intelligence fully and develop things that would be complementary with human intelligence rather than to create new intelligent species that would replace people, who are born with so many shortcomings, as claimed by some AI researchers who think that super-intelligent being is the natural result of cosmological evolution. 

The second point I’d like to comment on “one half of an argument” is that AI researchers need to open their mind to multiple approaches. Again, as a totally lay man, I think that “reverse engineering the human brain” could be called a bottom-up approach, which seems to be dominant approach in current AI. In Gerd Gigerenzer’s paper, Smart Heuristics, he cited a very interesting example about  how a baseball player catches a ball and he coined  an expression ,“gaze heuristic”. If AI researchers make more efforts to reveal how such heuristic works, then this kind of exploration could be called top-down approach. By combining both bottom-up and top-down approaches, AI field might make more progress and it may help to gain more friendly attitude towards AI from those skeptics who think the bottom-up only approach in AI is a mission impossible.

ISTIC is an institute under Ministry of Science and Technology of China (MOST). A main part of  ISTIC is National Engineering and Technology Library. In this sense, ISTIC is a bit like Library of Congress. On the other hand, we also provide decision-oriented information analysis service to officials of MOST. In that sense, ISTIC is like the Congressional Research Service. My work falls in the latter area.

Wish you all the success in your editing work!

Wu Yishan

ISTIC