Competition-level code generation with AlphaCode
Yujia Li, David Choi, Junyoung Chung et al. · Science · 2022
Programming is a powerful and ubiquitous problem-solving tool. Systems that can assist programmers or even generate programs themselves could make programming more productive and accessible. Recent transformer-based neural network models show impressive code generation abilities yet still perform poorly on more complex tasks requiring problem-solving skills, such as competitive programming problems. Here, we introduce AlphaCode, a system for code generation that achieved an average ranking in the top 54.3% in simulated evaluations on recent programming competitions on the Codeforces platform. AlphaCode solves problems by generating millions of diverse programs using specially trained transformer-based networks and then filtering and clustering those programs to a maximum of just 10 submissions. This result marks the first time an artificial intelligence system has performed competitively in programming competitions.
Cited by 1,016 works. doi:10.1126/science.abq1158 Open-access PDF
Authors
- Nate Kushman
- Julian Schrittwieser
- Rémi Leblond
- Tom Eccles
- James Keeling
- Felix Gimeno
- Agustin Dal Lago
- Thomas Hübert
- Cyprien de Masson d’Autume
- I. Babuschkin
- Xinyun Chen
- Po-Sen Huang
- Johannes Welbl
- Sven Gowal
References
- AI-Assisted Pipeline for Dynamic Generation of Trustworthy Health Supplement Content at Scale
- Scratch
- Toward automatic program synthesis
- SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing
- Grandmaster level in StarCraft II using multi-agent reinforcement learning
- CodeBERT: A Pre-Trained Model for Programming and Natural Languages
- Hierarchical Neural Story Generation
- Application of Theorem Proving to Problem Solving
- Understanding Back-Translation at Scale
- Automating string processing in spreadsheets using input-output examples