Abstract | ||
---|---|---|
We attacked the problem of solving crossword puzzles by computer: given a set of clues and a crossword grid, try to maximize the number of words correctly filled in. In our system, "expert modules" special- ize in solving specific types of clues, drawing on ideas from information retrieval, database search, and ma- chine learning. Each expert module generates a (pos- sibly empty) candidate list for each clue, and the lists are merged together and placed into the grid by a cen- tralized solver. We used a probabilistic representation throughout the system as a common interchange lan- guage between subsystems and to drive the search for an optimal solution. PROVERB, the complete system, averages 95.3% words correct and 98.1% letters correct in under 15 minutes per puzzle on a sample of 370 puz- zles taken from the New York Times and several other puzzle sources. This corresponds to missing roughly 3 words or 4 letters on a daily 15 x 15 puzzle, making PROVERB a better-than-average cruciverbalist (cross- word solver), |
Year | Venue | Keywords |
---|---|---|
1999 | National Conference on Artificial Intelligence | crossword solver,probabilistic cruciverbalist,centralized solver,puzzle source,expert module,database search,crossword puzzle,crossword grid,complete system,better-than-average cruciverbalist,new york times,information retrieval |
DocType | ISBN | Citations |
Conference | 0-262-51106-1 | 25 |
PageRank | References | Authors |
3.97 | 3 | 10 |
Name | Order | Citations | PageRank |
---|---|---|---|
Greg A. Keim | 1 | 87 | 12.95 |
Noam Shazeer | 2 | 1089 | 43.70 |
Michael L. Littman | 3 | 9798 | 961.84 |
Sushant Agarwal | 4 | 25 | 4.99 |
Catherine M. Cheves | 5 | 25 | 3.97 |
Joseph Fitzgerald | 6 | 25 | 3.97 |
Jason Grosland | 7 | 25 | 3.97 |
Fan Jiang | 8 | 59 | 12.93 |
Shannon Pollard | 9 | 25 | 3.97 |
Karl Weinmeister | 10 | 25 | 3.97 |