Ya, I never liked that abstract stuff that has nothing to do with real computers. It's why whenever I took one of those courses I dropped it shortly thereafter.
On Sat, Jan 26, 2013 at 5:58 AM, Jim Bromer <[email protected]> wrote: > Abram, > You idea that the situation could represent non-determinism is interesting > - especially since it fits in so nicely with the history of computation. > Turing developed his hypothesis of a non-deterministic (Turing) machine in > the 1930's. So your interpretation makes some historical sense (although > it is not what the professor of the course intended of course). If we were > to "diagram" the logic of a Turing Machine that was able to find a solution > you could use some kind of parallelism but you could also design a loop > that tried every possible action until it got a solution or which could > find more than one solution by examining every possible case. > > The other thing that bothered me was that since this is a AI Planning > course you can think of possible states as variables as well. I really > wonder if confusing common terminology with ways of thinking about a > problem drags students down or not. There are good reasons to present > students with historical issues, but why not just explain that is what you > are doing? > Jim Bromer > > On Fri, Jan 25, 2013 at 8:34 PM, Abram Demski <[email protected]>wrote: > >> Sounds annoying, sure. Perhaps it gets better? >> >> For what it's worth, I believe the idea with multiple copies of an action >> label is non-determinism. If there are several outgoing links with the same >> label, then the choice of next state is not fully determined by the action: >> it could be any of those. >> >> This can be used to represent randomness, but the idea of non-determinism >> in computer science is more general than that. It might be chosen according >> to some outside criteria. The most common case is nondeterminism in NP: NP >> indicates that there is a polynomial-time solution algorithm on >> nondeterministic turing machines, meaning Turing machines whose heads have >> multiple transition possibilities in their state transition diagrams, but >> "magically" make the correct transition to get the end result we need. >> >> >> On Fri, Jan 25, 2013 at 4:29 PM, Jim Bromer <[email protected]> wrote: >> >>> I am taking two Introduction to AI online courses and I am making quite >>> a few mistakes. Some of the mistakes are just plain mistakes. But some are >>> due to annoying cutsieness. Here is an example: >>> >>> True or False >>> State transition systems can represent actions that occur in parallel. >>> >>> Well of course they can, but since this is an introductory course the >>> answer has to be False. >>> Ok, got that one right. >>> >>> A state transition graph may have multiple outgoing/incoming directed >>> edges that are labelled with the same action. >>> >>> Well I had a feeling that was a trick question given the answer to the >>> first question, and the answer is...True. True? My best guess is that the >>> term label was not being applied to *some particular* labelled actions but >>> to the choice of actions at a particular state, whereas each state is the >>> resultant of the action so they are thought of as particulars (like >>> values)? Maybe there is some other reasoning behind this but if there is I >>> can't figure it out. >>> >>> This was from an introductory course in AI planning. AI Planning! I >>> find this stuff intensely annoying. We cannot use a state transition >>> diagram to diagram parallel actions? (That would be impossible for anyone >>> to even consider. Your mind can't handle it.) But it can be used to >>> represent multiple outgoing/incoming directed edges labelled with the same >>> action? Is the teacher kidding? Maybe you guys who have already been >>> through this see some error that I don't see, but I just don't see why the >>> teacher cannot just come out and explain the conventions that are applied >>> to terminology (like "state transition diagram") and save the quizzes for >>> the good stuff. Why challenge the students with the professor's mastery of >>> ambiguity? >>> >>> I make a lot of really dense mistakes in these courses. I don't want to >>> waste my time trying to outwit the teacher's challenges about computational >>> phlogiston. >>> >>> Jim Bromer >>> >>> >>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>> <https://www.listbox.com/member/archive/rss/303/7190161-766c6f07> | >>> Modify <https://www.listbox.com/member/?&;> Your Subscription >>> <http://www.listbox.com> >>> >> >> >> >> -- >> Abram Demski >> http://lo-tho.blogspot.com/ >> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >> <https://www.listbox.com/member/archive/rss/303/10561250-470149cf> | >> Modify <https://www.listbox.com/member/?&;> Your Subscription >> <http://www.listbox.com> >> > > *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> > <https://www.listbox.com/member/archive/rss/303/5037279-a88c7a6d> | > Modify<https://www.listbox.com/member/?&;>Your Subscription > <http://www.listbox.com> > ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
