It's different. Beam search isn't complete because it throws away candidates to focus on just a few, to try and save time and memory. I think the analogy is that it focuses like a beam.
>From the wikipedia page: "... But in beam search, only a predetermined number of best partial solutions are kept as candidates.[1] It is thus a greedy algorithm." On Wednesday, May 23, 2018 at 3:01:44 PM UTC-4, Amirouche Boubekki wrote: > > > > On Saturday, December 17, 2016 at 2:09:30 AM UTC+1, Nehal Patel wrote: >> >> A line by line account will be quite nice! >> >> I've had the pleasure of hearing several of your former students give >> tutorials on (non-relational) scheme interpreters. Without fail, they each >> recall with pure glee being shown how a mere apostrophe can separate a >> Lisp-2 from a Lisp-1... >> >> It feels there is a similar to be had opportunity in your exposition of >> microkanren: >> >> THIS line gives you "interleaving search" -- a mere transposition of >>> arguments separates it from a bottomless abyss (of depth first...) >> >> >> Preaching the gospels of interleaving search is one part of the 'kanren >> pedagogy (which is really one of best developed pedagogies in all of CS) >> that perhaps could be further emphasized. >> > > Isn't "interleaving search" another expression for "beam search" see > https://en.wikipedia.org/wiki/Beam_search > > >> In my own experience with porting microkanren -- trying to hunt down >> where the "magic lived" ended up being the most transformative part of the >> exercise. Perhaps a catchy phrase might help: "The unreasonable >> effectiveness of interleaving search for relational fixed points" (or some >> such nonsense) >> >> cheers, >> nehal >> > -- You received this message because you are subscribed to the Google Groups "minikanren" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/minikanren. For more options, visit https://groups.google.com/d/optout.
