cytron@uiuccsb.UUCP (09/19/83)
#N:uiuccsb:5500003:000:2322 uiuccsb!cytron Sep 18 19:12:00 1983 Forward at the request of speaker: /***** uiuccsb:net.arch / umcp-cs!speaker / 12:20 am Sep 17, 1983 */ The fact remains that if we don't have the algorithms for doing something with current hardware, we still won't be able to do it with faster or more powerful hardware. The fact remains that if we don't have any algorithms to start with then we shouldn't even be talking implementation. This sounds like a software engineer's solution anyway, "design the software and then find a CPU to run it on." New architectures, while not providing a direct solution to a lot of AI problems, provide the test-bed necessary for advanced AI research. That's why everyone wants to build these "amazingly massive" parallel architectures. Without them, AI research could grind to a standstill. To some extent these efforts change our way of thinking about problems, but for the most part they only speed up what we knew how to do already. Parallel computation is more than just "speeding things up." Some problems are better solved concurrently. My own belief is that the "missing link" to AI is a lot of deep thought and hard work, followed by VLSI implementation of algorithms that have (probably) been tested using conventional software running on conventional architectures. Gad...that's really provincial: "deep thought, hard work, followed by VLSI implementation." Are you willing to wait a millenia or two while your VAX grinds through the development and testing of a truly high-velocity AI system? If we can master knowledge representation and learning, we can begin to get away from programming by full analysis of every part of every algorithm needed for every task in a domain. That would speed up our progress more than new architectures. I agree. I also agree with you that hardware is not in itself a solution and that we need more thought put to the problems of building intelligent systems. What I am trying to point out, however, is that we need integrated hardware/software solutions. Highly parallel computer systems will become a necessity, not only for research but for implementation. - Speaker -- Full-Name: Speaker-To-Animals Csnet: speaker@umcp-cs Arpa: speaker.umcp-cs@UDel-Relay This must be hell...all I can see are flames... towering flames! /* ---------- */