Lecture 18-1 Adapting resolution for first-order logic — Transcript
Full transcript
- 0:00now let us study uh resolution proof
- 0:02system for first order logic we have
- 0:05seen reputation proof systems before
- 0:08let's remind ourselves what reputation
- 0:10proof systems are and how we can use
- 0:13resolution as a reputation proof system
- 0:15for our first order logic so we assume
- 0:17that we have a set of formulas in the
- 0:19left hand side in this uh statement
- 0:22which was treated as conjunction of
- 0:24formulas remember that right this is
- 0:26conjunction of these formulas should
- 0:28prove this Formula F
- 0:31so conjunction of CNF formula is also a
- 0:33CNF formula so if we have only CNF
- 0:36formula in the left hand side then we
- 0:39can treat as a one joint CNF formula
- 0:42implying other CNF formula that may
- 0:45simplify our view of the proving
- 0:49process let us suppose we ask to derive
- 0:52that Sigma proves F what do we do we
- 0:55assume Sigma is finite we can relax this
- 0:58condition due to compactness of first
- 1:00logic we will convert uh conjunction of
- 1:04Sigma and not of F and turn into a first
- 1:09order Clauses called Sigma
- 1:12Prime and we apply the reputation proof
- 1:15system on Sigma Prime if it gives us
- 1:18false then this original statement was
- 1:24derivable now now how we build a
- 1:27resolution proof system for our Sigma
- 1:29Prime so there are two steps we needed
- 1:32one resolution which very much like
- 1:33proposition logic and new thing has come
- 1:35up with unification which we have seen
- 1:37in the previous
- 1:38lecture okay so what was resolution
- 1:41resolution was like this okay so you
- 1:43have F or C and not F or D you can
- 1:46derive C or D this F and not F get
- 1:48cancelled and then Union is applied yeah
- 1:51so that is the sense of resolution
- 1:54however there is a complication first
- 1:56order logic you can have variables in
- 1:59your first order logic object and these
- 2:00variables may not be same across Clauses
- 2:03so
- 2:04therefore you may not be immediately
- 2:06able to apply
- 2:08the
- 2:11resolution for example you have these
- 2:13two Clauses in which you have not p z z
- 2:18and PX FY okay so you can see that these
- 2:22two Clauses are two different
- 2:25literals however they are unifiable
- 2:28literals yes so X can be mapped to zed
- 2:31and the FX can be mapped to zed and then
- 2:34this whole unification can result you as
- 2:36some some mapping between variable XY Z
- 2:39and then you can make the mean
- 2:41call so uh if you can unify using
- 2:46unification then we can apply the
- 2:47resolution the now the question comes
- 2:51how do we
- 2:52apply uh unification in our Clauses okay
- 2:56so we were unifying terms earlier now
- 2:59I'm unifying atoms that looks weird so
- 3:02let's we will think about
- 3:04it uh is substitution a valid operation
- 3:08for derivations I mean like you know if
- 3:11we have some clause which is for all VAR
- 3:14you know all the variables are
- 3:15universally Quantified suddenly you take
- 3:17a variable replaced by another term is
- 3:19it allowed yeah and not only one
- 3:21variable you're doing from multiple
- 3:23variables in the same time is it an okay
- 3:26derivation so that we need to think
- 3:28about
- 3:30the third point is that uh what is the
- 3:33meaning of the variables across Clauses
- 3:36if same variable appears in two
- 3:37different Clauses what does it mean are
- 3:39the same variable or the different
- 3:40variables uh how their unification works
- 3:43so we need to think about this third
- 3:44matter
- 3:46also okay so first let's take the issue
- 3:49one by one unification of atoms we can
- 3:52lift the idea of unifying terms to
- 3:55unifying atoms simply treat a predicate
- 3:58as a function and then you are set you
- 4:00can just apply any unification algorithm
- 4:03for example if you want to unify this
- 4:05guy to this guy you simply uh what do
- 4:08you do you say okay if x should be equal
- 4:10to this guy this guy should be equal to
- 4:12this guy can be unify okay and if you do
- 4:15that then you'll get this map X should
- 4:17be M to FY Z should be ma to FY and then
- 4:20you are
- 4:22done after applying this unification we
- 4:25obtain this uh this terms
- 4:30now now let's look at the second issue
- 4:32so so we will go by an example okay so
- 4:35we know that the following derivation is
- 4:37valid okay what is this following
- 4:38deration let's suppose I give you for
- 4:40all XY FX y we can replace each letter
- 4:45by a term let say T1 XY T2 XY okay and
- 4:50then we can apply un Universal
- 4:52quantification again this was given to
- 4:54you as an exercise in previous lectures
- 4:57you can check that these two proofs
- 4:59steps are satisfying the conditions side
- 5:05conditions given to you in the earlier
- 5:07lectures therefore the following
- 5:10derivation in our Clauses are some I
- 5:13give you a clause it has implicit
- 5:15Universal quantifiers remember that we
- 5:16don't write Universal quantifiers and
- 5:18you simply apply a substitution the
- 5:20resulting thing is also okay so
- 5:23therefore uh you don't need to worry
- 5:25about what substitution is allowed what
- 5:26naming is allowed doesn't matter you
- 5:28just apply any substitution it is going
- 5:30to go work it is going to
- 5:32work the following deration is a valid
- 5:36derivation let's suppose I give you for
- 5:38all PX or qy and then the sigma is y
- 5:43maps to X so I can simply say PX or QX
- 5:46this isct so implicitly there was a
- 5:49universal quantifier XY now this is for
- 5:52all X okay so just keep that remember
- 5:55remember that there's a universal
- 5:56Quantified setting in front of all the
- 5:58clauses
- 6:00the third issue the variables across the
- 6:02Clauses how do we make sense out of them
- 6:04okay the universal quantifi distribute
- 6:06over conjunction that was the fact so if
- 6:08so we can easily distribute the
- 6:10quantifiers and scope only each Clauses
- 6:13so no quantifi scopes across Clauses so
- 6:16the variables appearing across Clauses
- 6:18are not the same variables even if we
- 6:20write the same name for example if I
- 6:22give you this uh formula uh it's wrong
- 6:25way to write it you should just simply
- 6:27distribute it okay so you get this
- 6:30distributed form in which you have a w
- 6:32appearing twice okay and you can say
- 6:36that you know they are different
- 6:38variables right so therefore we may view
- 6:41the variables occurring in different
- 6:42Clauses as different variables even if
- 6:44we use the same name because if you each
- 6:46time you start writing different name
- 6:48you run out of the name whole notation
- 6:49becomes very cumbersome uh human
- 6:52reability goes down okay so therefore
- 6:55for humanability we reuse the same name
- 6:58but uh mically or when computers analyze
- 7:01these these Clauses they treat them
- 7:06disparately so please pay attention it's
- 7:08a big source of confusion okay so when
- 7:10you're doing unification and trying to
- 7:12do reputation WR reputation proofs this
- 7:14can confuse you
About this transcript
This page contains the full transcript of Lecture 18-1 Adapting resolution for first-order logic by Automated Reasoning, generated from the public captions YouTube serves with the video. The transcript has 1,152 words across 170 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.