Automated Crossword Solving

  title={Automated Crossword Solving},
  author={Eric Wallace and Nicholas Tomlin and Albert Xu and Kevin Yang and Eshaan Pathak and Matthew Ginsberg and Dan Klein},
We present the Berkeley Crossword Solver, a state-of-the-art approach for automatically solving crossword puzzles. Our system works by generating answer candidates for each crossword clue using neural question answering models and then combines loopy belief propagation with local search to find full puzzle solutions. Compared to existing approaches, our system improves exact puzzle accuracy from 57% to 82% on crosswords from The New York Times and obtains 99.9% letter accuracy on themeless… 



Learning to Rank Answer Candidates for Automatic Resolution of Crossword Puzzles

It is shown that learning to rank models based on relational syntactic structures defined between the clues and the answer can improve both modules above and improve the resolution accuracy of crossword puzzles.

Distributional Neural Networks for Automatic Resolution of Crossword Puzzles

This paper creates a labelled dataset of 2 million clues on which an innovative Distributional Neural Network (DNN) is applied for reranking clue pairs and shows a large improvement over the TK approach, when the latter uses small training data.

WebCrow: A Web-Based System for Crossword Solving

An overview of WebCrow, a system that tackles crosswords using the Web as a knowledge base that appears to be a novel approach with respect to the available literature, which is the first solver for non-English crosswords and has been designed to be potentially multilingual.

Dr.Fill: Crosswords and an Implemented Solver for Singly Weighted CSPs

Dr. Fill is described, a program that solves American-style crossword puzzles by converting crosswords to weighted csps, and then using a variety of novel techniques to find a solution.

Quizbowl: The Case for Incremental Question Answering

This work makes two key contributions to machine learning research through Quizbowl: collecting and curating a large factoid QA dataset and an accompanying gameplay dataset, and developing a computational approach to playing Quiz Bowl that involves determining both what to answer and when to answer.

On NMT Search Errors and Model Errors: Cat Got Your Tongue?

It is concluded that vanilla NMT in its current form requires just the right amount of beam search errors, which, from a modelling perspective, is a highly unsatisfactory conclusion indeed, as the model often prefers an empty translation.

Know What You Don’t Know: Unanswerable Questions for SQuAD

SQuadRUn is a new dataset that combines the existing Stanford Question Answering Dataset (SQuAD) with over 50,000 unanswerable questions written adversarially by crowdworkers to look similar to answerable ones.

Decrypting Cryptic Crosswords: Semantically Complex Wordplay Puzzles as a Target for NLP

A novel curriculum approach, in which the model is first fine-tuned on related tasks such as unscrambling words, and investigates model systematicity by perturbing the wordplay part of clues, showing that T5 exhibits behavior partially consistent with human solving strategies.