Artificial Intelligence: Achievements
- Updated

On this page
Learning and search at professional Go level
Demonstrated result · March 2016
AlphaGo defeated Lee Sedol by four games to one. Combining learned evaluations with search was a major demonstration in a demanding strategic game. The achievement concerns Go under defined rules; it does not demonstrate unrestricted reasoning in the physical world.
AlphaGo research and the 2016 match — Google DeepMind · Checked
Predicting biomolecular interactions
Demonstrated result · 8 May 2024 announcement
AlphaFold 3 extended structure prediction to complexes involving proteins and other molecules. This is an important computational contribution to biological research. A predicted structure is a model to investigate, not an experimentally established interaction or proof that a proposed medicine is safe and effective.
AlphaFold research programme and AlphaFold 3 announcement — Google DeepMind · Checked