DeepMind AlphaCode Achieves Median Human Level in Coding Competitions
DeepMind's AlphaCode system writes computer programs at a competitive level, ranking in the top 54% of participants on Codeforces. The tool uses massive-scale code generation and smart filtering to solve novel problems.
DeepMind introduces AlphaCode, an artificial intelligence system that writes computer programs at a competitive level. By achieving an estimated rank within the top 54% of participants on Codeforces, the system sets a new milestone for AI in complex problem-solving tasks.
The system relies on transformer-based language models to generate code at an unprecedented scale and then intelligently filters the output down to a small set of promising solutions. This approach allows AlphaCode to tackle novel problems that require a blend of critical thinking, logic, algorithms, coding, and natural language understanding.
Validated across 10 recent programming contests that are newer than its training data, AlphaCode performs at about the level of the median competitor. To encourage further innovation in the field, DeepMind releases a comprehensive dataset of competitive programming problems and solutions on GitHub.