February 11, 1985
Comrade Hong Jiawei, Maxi Wen comrades:
The last time you two talk, Comrade Maxi Wen comes to the development of parallel computing supercomputer machine intelligence will be useful, instructive for me, I've been thinking. Now see, Comrade Greek text articles in "philosophy" in 1985, one more thought 1981 5 article, it seems to have income, set out below, to ask:
I used to know the logic is monotonic logic, no matter how long the reasoning process, are presented linear, access is not on other reasoning string. So I once logical thinking (that is, abstract thought?) Is called linear. Now we know there is nonmonotonic logic inference may be attached to a line from another line, cross, then into the net. This is not a linear extension of the plane, "one-dimensional" to "two-dimensional." I had once image (Xenia) thinking called surface type.
Thus, the formation of a reasoning system, but this system is not a polycarboxylic or polyol algebraic simultaneous equations to express the simultaneous differential equations, but rather to express a logic symbol. The algebraic expression systems using multiple simultaneous equations or multivariate simultaneous differential equations can exhibit many special features, this is what I said systematics (H.Haken called "collaborative learning", synergeha); then the logical expression symbol the system is also likely to exhibit the kind of monotonous logical line of thinking does not. It may be smart, be direct sense, is to create!
We study intelligent, this is the key! Fuzzy reasoning is not necessarily true of the way! We want to establish a logical systematics. (Point of view of this system, artificial intelligence has also been working in the signs).
Above ideas may be nonsense, please two exhibitions.
Sincerely,
salute
Qian
1985.2.11
Please refer to my book 238,263 pages
Annotation: H.Haken: Hermann Haken, Hermann Haken (1927-), German physicist, founder of Synergetics.
June 6, 1985
Professor Hong Jiawei:
When you see this letter, a series of conferences China Computer Federation probably come to an end. I am very disappointed with our performance in the computer industry zhuoxian meeting: the lack of a clear guiding ideology, China's computer industry really worrying!
The core is the lack of talented theoretical guidance: At the end of sixty of computer industry because it can not see the role of software technology, and mistakes; today it? Or behind a large portion of people while I do not feel!
So I think it must grasp the theory, there must be a guiding theory; theory to guide practice, practice ≠ real theory! The current theory of intelligent machines - scientific thinking, must grasp go.
In this regard you have done a very good job. But I'm not satisfied, because you talk about the upper limit of the insurance too. We are looking for tips. Is your work a bit like Klee and Minty the theory of linear programming problems, the most strange, the most biased situation as the theme, so that is the NP. I've suggested the network with Synergetics, but also simulate the human brain, looking for tips Well! Why do not you consider the network? (Read Jin Guantao: "logical paradox and self-organizing systems", "Dialectics of Nature" 1985, 2, 7).
Theory must serve for the actual build intelligent machines. Not only for the papers.
How do you think?
Sincerely,
salute!
Qian 1985.6.6
Annotation: NP: Nonlinear Programming, translated as "non-linear programming."