Nicht angemeldet. · Kennwort vergessen · Registrieren

Seite:  1  2  nächste
 Mitglied seit 04/2018 15 Beiträge 04.02.2019, 22:15   #1   Betreff: More Training Data for Exam Preparation [Logic] +1 mecanum I wanted to kindly ask whether it would be possible to provide us with more example logic problems, that are accompanied by solutions in the lecture's notation - specifically the kind which is similar to the following: Problem 10.1 (Calculi Comparison) Problem 11.2 (First-Order Semantics) Problem 11.3 (Natural Deduction) Problem 12.2 (First-Order Resolution) Problem 12.3 (First-Order Tableaux) At least for me, access to more (and diverse examples) would greatly contribute to getting a finer intuition for the subtle differences between the used calculi. Edit: Also super appreciation is offered in exchange for the solution for Assignment 12 Dieser Beitrag wurde am 04.02.2019, 22:48 von lu60ruhy verändert.
 KI 1 Tutor WS18/19 Mitglied seit 10/2014 93 Beiträge 05.02.2019, 07:26   #2   +5 lu60ruhy, Jonas S, Jazzpirate, Marcel[Inf], LasagneAlForno >makes a post demanding for Data >includes Star Trek Character which is not Data 0/10 meme potential used No, honestly, we will see if we have/find some more logic problems which are not used in your exam... "Debugging is like doing surgery by randomly squeezing stuff in a patient's body and going like 'lmao tell me when this guy stops breathing'." -- Orteil, Creator of Cookie-Clicker
 Mitglied seit 10/2016 743 Beiträge 05.02.2019, 12:06   #3   Antwort auf Beitrag #1 +1 lu60ruhy I wanted to kindly ask whether it would be possible to provide us with more example logic problems You can easily generate infinitely many of them Any valid formula is a good example for resolution, natural deduction and tableau, any non-valid formula is a good example for counter-model generation using tableau. Recall my previous post on algebraically handling formulas - you can easily use them to generated arbitrarily complex tautologies. https://fsi.cs.fau.de/forum/thread/16935-Exercise-10-1-Tip… Another fun (and educational!) thing you can do is: Generate interesting formulas by running calculi backwards. For example, to generate an interesting tautology, start with an empty clause and "invent" clauses that resolve to the empty clause. Then invent clauses that resolve to the resulting clauses etc. Try to do non-trivial things - what do clauses have to look like such that they resolve to the empty (or previous) clauses and I need to unify terms? What if I have a skolem function here, what quantifiers does the resulting formula then have? Then try to prove that formula using a *different* calculus Use natural deduction without a specific goal to find a tautology, then prove it using tableau etc. All of this will not just give you infinitely many exercises for practicing calculi, but running things backwards will additionally give you a better intuition how any why that calculus works, what kinds of mistake you can make ("Wait, by running resolution backwards I seem to have generated a formula that isn't valid - where did I go wrong and why?") and as a result, you will be a lot faster in the exam
 Mitglied seit 10/2016 743 Beiträge 05.02.2019, 12:10   #4   +1 lu60ruhy Furthermore: Find a non-trivial formula "for each rule" in a calculus. If there's a side condition (e.g. "x may not occur freely in an open premise"), find situations where they don't apply and why applying the rule anyway would be a problem. All of this is a lot more educational than If I were to just hand you formulas to mechanically prove - and counterintuitively, you will become better at proving things faster by going about it that way than by merely practicing proving things Don't just read it; fight it! Ask your own questions, look for your own examples, discover your own proofs. Is the hypothesis necessary? Is the converse true? What happens in the classical special case? What about the degenerate cases? Where does the proof use the hypothesis? --- Paul R. Halmos Dieser Beitrag wurde am 05.02.2019, 12:29 von Jazzpirate verändert.
 Mitglied seit 04/2018 15 Beiträge 05.02.2019, 12:48   #5   Thank you for your answer! I'm afraid I have to disagree with you. If it were true, that your proposed way of generating the problems by one's self is the best way to gain an intuition into the calculi, you would probably have presented it this way in the exercises. I'm sure that it is a good way to deepen the understanding but at least I am not there yet. The fact that every additional source of similar problems outside the exercises uses a slightly different notation / terminology complicates grasping the essence of the calculi. As far as I can tell [RN] does not offer these kind of problems in the exercise part as well (maybe I am mistaken?). So before I can make the step to speeding up, I would very much like to verify my mental model of the calculi by training on examples that are accompanied by solutions with homogeneous notation that I can be confident in (contrary to solutions, which might be incorrect, because I 'invented' them myself - solving/running them in either direction). tl;dr: Let me get comfortable with the mechanical stuff first Thank you for your effort! Dieser Beitrag wurde am 05.02.2019, 12:53 von lu60ruhy verändert.
 Mitglied seit 10/2016 743 Beiträge 05.02.2019, 13:03   #6   +1 lu60ruhy If it were true, that your proposed way of generating the problems by one's self is the best way to gain an intuition into the calculi, you would probably have presented it this way in the exercises. Except that it's difficult to formualte and evaluate an exercise that basically amounts to "think for yourself" Exercises need to be self-contained with an (ideally) clear instruction what you're expected to do and how the result will be graded. I would very much like to verify my mental model of the calculi by training on examples that are accompanied by solutions with homogeneous notation that I can be confident in Fair point I'll see what I can do, but I have to admit that I'm pretty swamped at the moment, and nicely presenting formulas *and* their proofs in a given calculus takes a lot of time. I'll happily look over solutions that you or your coeds post in the forum, though. Practice, Practice & More PracticePerfectly correct - "Practice" goes beyond mere mechanics though. Sure, you can't fully understand a calculus without practicing it, but you can't understand a calculus by merely mechanically executing the algorithm over and over again either. Just as you can't become a good piano player without doing finger exercises, merely doing finger exercises won't make you a good piano player.
 Mitglied seit 04/2018 15 Beiträge 05.02.2019, 14:58   #7   I'm pretty swamped at the moment, and nicely presenting formulas *and* their proofs in a given calculus takes a lot of time.Maybe the tutors can help idk what they actually signed up for ... In the mean time at least the solution to Assignment 12 would be beneficial
 Mitglied seit 10/2016 743 Beiträge 05.02.2019, 15:46   #8   +1 lu60ruhy
 Mitglied seit 07/2016 79 Beiträge 05.02.2019, 21:48   #9   You might find https://fsi.cs.fau.de/dw/pruefungen/bachelor#grundlagen_de… helpful.
 Mitglied seit 04/2018 15 Beiträge 06.02.2019, 16:28   #10   +1 Nash Thanks for the advice! I had already looked into that, but unless I am missing something, the only accessible solutions there are a.) in a slightly different notation and b.) created by students (titled "solution attempts") which does not induce a lot of confidence (which is not a big deal, unless you are trying to learn this from scratch ) I know this sounds picky but I am convinced that these things matter when you are new to something. Consistency is king and errors in introductory material (or something that is used as such) pose considerable throwbacks when the student is trying to make sense of a concept. And I know that everybody (who puts in their time) will manage without the "extra training data" I asked for - but I am convinced, that it would help a lot of people get there faster. And thank you again for putting in thought and work (as much as it is possible - I know you have a lot to do elsewhere)! Dieser Beitrag wurde am 06.02.2019, 16:51 von lu60ruhy verändert.
 Mitglied seit 11/2018 1 Beitrag 06.02.2019, 18:22   #11   Antwort auf Beitrag #1 +1 lu60ruhy Zitat von lu60ruhy am 04.02.2019, 22:15:At least for me, access to more (and diverse examples) would greatly contribute to getting a finer intuition for the subtle differences between the used calculi. I fully agree with you. I am also looking for some more exercises (especially for natural deduction in Gentzen notation)  and reliable solutions. It is hard to figure out, how to proceed, if there are just a few examples. Every kind of exercise/ solution would be helpful for me.
 Mitglied seit 04/2018 15 Beiträge 06.02.2019, 22:12   #12   Is the solution to Problem 5.2 (First-Order Tableau) (https://new.kwarc.info/teaching/AI/exam-WS2017-18-with-sol…) intentionally kept secret?
 Mitglied seit 10/2016 743 Beiträge 07.02.2019, 12:21   #13   I'll do it on a whiteboard and send you a photo, okay?
 Mitglied seit 04/2018 15 Beiträge 07.02.2019, 13:02   #14   A voice message would have been better, but if you insist ...
 Mitglied seit 10/2016 743 Beiträge 07.02.2019, 13:51   #15   +1 lu60ruhy
Schließen Kleiner – Größer + Auf diesen Beitrag antworten:
Prüfcode: Gib bitte das Wort aus dem Bild ins folgende Textfeld ein. (Nur die Buchstaben eingeben, Kleinschreibung ist in Ordnung.)
Smileys:
Weitere Zeichen:
Seite:  1  2  nächste
Gehe zu Forum
Datenschutz | Kontakt