An AI a Day Keeps Dr.Fill at Play: Matt Ginsberg on Building GPU-Powered Crossword Solver

by Clarissa Garza

9 Down, 14 letters: Someone skilled in creating and solving crossword puzzles.

This April, the fastest “cruciverbalist” at the ​​American Crossword Puzzle Tournament was Dr.Fill, a crossword puzzle-solving AI program created by Matt Ginsberg.

Dr.Fill perfectly solved the championship puzzle in 49 seconds. The first human champion, Tyler Hinman, filled the 15×15 crossword in exactly three minutes.

Though Ginsberg has published crossword puzzles for the New York Times, he has trouble solving puzzles, even his own. After attending a crossword tournament over a decade ago, Ginsberg decided to create a crossword-solving program to compete against top-tier word nerds.

Ginsberg spoke with NVIDIA AI Podcast host Noah Kravitz about his decade-long journey creating Dr.Fill and where he envisions it going in the future.

Key Points From This Episode:

  • Ginsberg partnered with UC Berkeley’s natural language processing team. By combining their code bases, Dr.Fill outperformed the human competitors at this year’s tournament.
  • Dr.Fill performs better on unthemed rather than themed puzzles. This is where the natural language group comes into play, helping Dr.Fill interpret clues and difficult crossword themes.

Tweetables:

“I’m looking forward to the crossword tournament next year because I know we’re going to be working hard to make the program better.” — Matt Ginsberg [11:02]

You Might Also Like:

Otter.ai CEO Sam Liang on Bringing Live Captions to a Meeting Near You

Sam Liang is making things easier for the creators of the NVIDIA AI Podcast — and just about every remote worker. He’s the CEO and co-founder of Otter.ai, which uses AI to produce speech-to-text transcriptions in real time or from recording uploads.

How Deep Learning Can Translate American Sign Language

Rochester Institute of Technology computer engineering major Syed Ahmed, a research assistant at the National Technical Institute for the Deaf, uses AI to translate between American Sign Language and English. Ahmed trained his algorithm on 1,700 sign language videos.

Pod Squad: Descript Uses AI to Make Managing Podcasts Quicker, Easier

Serial entrepreneur Andrew Mason is making podcast editing easier and more collaborative with his company, Descript Podcast Studio, which uses AI, natural language processing and automatic speech synthesis.

Subscribe to the AI Podcast

Get the AI Podcast through iTunes, Google Podcasts, Google Play, Castbox, DoggCatcher, Overcast, PlayerFM, Pocket Casts, Podbay, PodBean, PodCruncher, PodKicker, Soundcloud, Spotify, Stitcher and TuneIn. If your favorite isn’t listed here, drop us a note.

Tune in to the Apple Podcast Tune in to the Google Podcast Tune in to the Spotify Podcast

Make the AI Podcast Better

Have a few minutes to spare? Fill out this listener survey. Your answers will help us make a better podcast.