Title: From Imitation to Discovery Test: A case study of AlphaGeometry.

Speaker: Thang Luong, Google DeepMind

​

Abstract: In this talk, I will discuss our recent work, AlphaGeometry, a neuro-symbolic system that can solve geometry at the Olympiad level. Key to our work is the idea of synthetic data generation at scale, in which we synthesize 100M theorems and proofs for AlphaGeometry to learn entirely from scratch. AlphaGeometry successfully solved 25 out of 30 recent Olympiad-level problems, surpassing previous methods and nearing the capabilities of an International Mathematical Olympiad (IMO) gold medalist. During the talk, I will also highlight the strengths and weaknesses of AlphaGeometry and hint towards a bigger picture of advancing reasoning capabilities of existing AI systems.

Bio: Thang Luong is currently a Senior Staff Research Scientist at Google DeepMind, ex Google Brain. He obtained his PhD in Computer Science from Stanford University in 2016, during which he pioneered the field of deep learning for machine translation. At Google, Dr. Luong built state-of-the-art models in both language (QANet, ELECTRA) and vision (UDA, NoisyStudent). He co-founded Project Meena, which debuted the world’s best chatbot in 2020 and later became Google LaMDA chatbot in 2021. Dr. Luong has been co-leading the development of Bard Multimodality since 2022 and is the principal investigator of the AlphaGeometry project (Nature, 2024) that solves Olympiad geometry problems at the IMO level.

​

​