This makes me curious, how are crossword puzzles made? Clues -> words? Or working backwards from words to clues? Some blend of the two?
What a crossword AI reveals about humans' way with words
31–40 of 50 posts
Re: What a crossword AI reveals about humans' way with words
#32Earlier quoted context omitted.
> Non-existent words should be easy to check for. I'm interested in your magical solution because this is a hard problem.
You should have given more detail on why this problem is deceptively hard. I am guessing that the simple solution of looking up the word in the dictionary seems to work ok (especially in the context of an artificial competition, which doesn't have to accept uncommon spellings, words in other languages etc), but still breaks down hard because of proper names, which are common in cross-words.
There is no dictionary that contains "all existing words" for any giving language.
Re: What a crossword AI reveals about humans' way with words
#33Earlier quoted context omitted.
> Non-existent words should be easy to check for. I'm interested in your magical solution because this is a hard problem.
You should have given more detail on why this problem is deceptively hard. I am guessing that the simple solution of looking up the word in the dictionary seems to work ok (especially in the context of an artificial competition, which doesn't have to accept uncommon spellings, words in other languages etc), but still breaks down hard because of proper names, which are common in cross-words.
I would definitely be interested in knowing why it is hard. And yes, I was thinking about lookup: for example from (1) a dictionary, (2) a set of proper names and (3) a set of previous crossword answers. While nothing is perfect, but this (from an armchair) seems like it should work pretty well. And I am not proposing it as the main part of the algorithm, just a check on non-words.
Re: What a crossword AI reveals about humans' way with words
#34This makes me curious, how are crossword puzzles made? Clues -> words? Or working backwards from words to clues? Some blend of the two?
The crossword writer I've seen discuss it, generally start with theme words, then fill in the rest of the words, then work on clues.
Then an editor will refine the clues to fit whatever criteria they may have for publication.
Re: What a crossword AI reveals about humans' way with words
#35Earlier quoted context omitted.
Actually, one of the interesting things about cryptic crosswords is that, in some ways, the rules are far more codified. The specifics definitely depend on the paper, of course, but basics like the fact that the majority of the clues are actually two halves, clueing the word two different ways, and that anagrams are always explicitly clues using words such as "scrambled," "drunken," "off" or whatever could actually m…
well yes, but the beautiful thing about UK cryptics at least is that they regularly intentionally misdirect the user, so you spend some time going down a rabbit hole and realise that you're on a hiding to nothing. I think that kind of game play is what makes a rules-based approach more challenging. I'll have to check out Crossword Genius though.
That might make it easier, not harder, for computers.
Adding in multiple distractor possibilities is a good way to make things harder for humans, with our limited working memory & difficulty chasing a lot of possibilities in parallel, but for a computer that's borderline trivial while it's evaluating each candidate accurately that's hard for a computer to get right. (Think of chess or Go where a computer can descend the game tree down millions of paths easily, but it's accurately evaluating a single board position which is extremely difficult; human players, on the other hand, can evaluate a board at a glance, but can only think through a few lines of play at a time.)
If the rules are simple but there's multiple candidate, the AI will just split at each possible candidate and pursue them until the AI can check whether it fits in the crossword grid or not.
Re: What a crossword AI reveals about humans' way with words
#36Earlier quoted context omitted.
I think it's pretty unlikely you'll be able to learn the UK rules to a human level using brute force ML - take a fairly easy clue like ‘Drunken men are more despicable' - you need to split this as (wordplay = drunken (anagram of) "men are") / (meaning = "more despicable") = "MEANER" This is really hard for supervised learning - the reward is quite sparse (e.g. did you get it right / how many characters did you get ri…
You might not need to codify the rules; you could create a tagged training set which includes additional information, like a ‘parts of speech’ breakdown of how a clue relates to an answer (anagramSignifier - anagramMaterial - association, etc.) An unsupervised learner could probably do reasonably well at picking up on those patterns, even to the point of considering that a word it hasn’t seen used as an anagram signi…
That's still codifying the rules - doing it through the data rather than in preprocessing
Re: What a crossword AI reveals about humans' way with words
#37"But, unlike more than 200 human solvers, it wasn’t perfect on all of the puzzles: It got waylaid on two of them and finished with errors." I'm curious, what kind of error are we talking about? Words that don't exist, or another solution to a problem that may not have a unique solution?
I wonder how it does on puzzles where the answers need to be written in unusual ways? Some examples from the New York Times puzzles. - There was one with a name that suggested an Alice in Wonderland connection, and it had an answer "THE LOOKING GLASS" (no spaces) running vertically down the full length of the center of the grid. Every across answer that was entirely to the left of that was written normally. Every acr…
Those are great. Another memorable NYT one had the theme A SHOT IN THE DARK, and several answers ended in SHOT (RIMSHOT, EARSHOT, etc.), but the grid only had space for the first part of the word, so the actual answers were RIM⬛, EAR⬛, etc. Another had a RISE FROM THE ASHES theme, and had several answers that ended in ASH that did not seem to fit the clue: the actual answer followed vertically where the ASH ending started. E.g., for the clue "Tell it like it is", the answer that fit straight across was "TALKS TRASH", but above the "A", going up, were I, G, H, and T, giving the actual answer "TALK STRAIGHT".
Re: What a crossword AI reveals about humans' way with words
#38A few years back I wrote a program that, given a blank or partially-filled in (NYT-style) crossword grid (with black squares already inserted), could fill out the rest of the grid with valid words/phrases both across and down. It had no relation to clues though, you had to write the clues yourself after the grid was filled out.
I wrote it because I wanted to make my dad (a huge nyt crossword fan) a custom crossword for his birthday. I put in a bunch of phrases related to him and our family and let the program fill in the rest. It was a huge hit, never really went back to it though. Anyone know if anything else like this exists?
Re: What a crossword AI reveals about humans' way with words
#39"But, unlike more than 200 human solvers, it wasn’t perfect on all of the puzzles: It got waylaid on two of them and finished with errors." I'm curious, what kind of error are we talking about? Words that don't exist, or another solution to a problem that may not have a unique solution?
A similar thing happens even with human solvers from time to time. It's possible to solve a puzzle with answers that match the clue but aren't what was intended. Of course, the more you let yourself deviate from the clue, the more possibilities exist.
My favorite example of using this is the 1996-11-05, the day of the presidential election, NY Times crossword with the clue "Lead story in tomorrow's newspaper (!), with 43-Across". The puzzle worked with both "CLINTON" or "BOBDOLE" in the crossword.
Re: What a crossword AI reveals about humans' way with words
#40This makes me curious, how are crossword puzzles made? Clues -> words? Or working backwards from words to clues? Some blend of the two?
There was a video next to the article for me that discussed exactly that ... https://youtu.be/aAqQnXHd7qk