Solving olympiad geometry without human demonstrations
By Trieu H. Trinh, Yuhuai Wu, Quoc V. Le, He He, Thang Luong. 2024
808 citations recorded by the source corpus.
Proving mathematical theorems at the olympiad level represents a notable milestone in human-level automated reasoning1–4, owing to their reputed difficulty among the world’s best talents in pre-university mathematics. Current machine-learning approaches, however, are not applicable to most mathematical domains owing to the high cost of translating human proofs into machine-verifiable format. The problem is even worse for geometry because of its unique translation challenges1,5, resulting in severe scarcity of training data. We propose AlphaGeometry, a theorem prover for Euclidean plane geometry that sidesteps the need for human demonstrations by synthesizing millions of theorems and proofs across different levels of complexity. AlphaGeometry is a neuro-symbolic system that uses a neural language model, trained from scratch on our large-scale synthetic data, to guide a symbolic deduction engine through infinite branching points in challenging problems. On a test set of 30 latest olympiad-level problems, AlphaGeometry solves 25, outperforming the previous best method that only solves ten problems and approaching the performance of an average International Mathematical Olympiad (IMO) gold medallist. Notably, AlphaGeometry produces human-readable proofs, solves all geometry problems in the IMO 2000 and 2015 under human expert evaluation and discovers a generalized version of a translated IMO theorem in 2004. A new neuro-symbolic theorem prover for Euclidean plane geometry trained from scratch on millions of synthesized theorems and proofs outperforms the previous best method and reaches the performance of an olympiad gold medallist.
Subfields: Artificial Intelligence, Information Systems, Reinforcement Learning, Program Analysis & Verification
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
In Information Systems
In Reinforcement Learning
In Program Analysis & Verification
Citations in the corpus
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models (2024, background): “Large language models (LLM) have revolutionized the approach to mathematical reasoning in artificial intelligence, spurring significant advancements in both the quantitative reasoning benchmark (Hendrycks et al., 2021) and the geometry reasoning benchmark (Trinh et al., 2024).”
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning (2025, neutral): “…various approaches, including process-based reward models (Lightman et al., 2023; Uesato et al., 2022; Wang et al., 2023), reinforcement learning (Kumar et al., 2024), and search algorithms such as Monte Carlo Tree Search and Beam Search (Feng et al., 2024; Trinh et al., 2024; Xin et al., 2024).”
- Chain of Thought Empowers Transformers to Solve Inherently Serial Problems (2024, background): “…final answers, through prompting (Wei et al., 2022; Chowdhery et al., 2022; Lampinen et al., 2022; Kojima et al., 2022; Zhou et al., 2023) or finetuning (Cobbe et al., 2021; Nye et al., 2021; Chung et al., 2022; Huang et al., 2022), dramatically enhances the performance of large language models…”
- AlphaEvolve: A coding agent for scientific and algorithmic discovery (2025, neutral): “…the FunSearch approach [24, 83] established LLM-guided evolution as a powerful tool for discovering witnesses for, and counterexamples to, mathematical statements—a problem that is complementary to that of finding formal and informal proofs of mathematical statements [3, 19, 98, 99, 110, 111].”
- OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems (2024, methodology): “…computational derivations, and was basically unable to complete the proof within the token limitation in Math-En_COMP, indicating that existing models still cannot effectively solve lengthy reasoning and proofs, which is consistent with the conclusions in existing papers (Trinh et al., 2024).”
- Large Language Monkeys: Scaling Inference Compute with Repeated Sampling (2024, neutral): “Similar tree-based methods can also be effective in combination with LLMs, allowing models to better plan and explore different approaches [63, 8, 53, 54].”
- Will we run out of data? Limits of LLM scaling based on human-generated data (2022, methodology): “26 For example, AlphaZero (Silver et al., 2017) was famously trained using self-play, and more recently Al-phaGeometry (Trinh et al., 2024) was trained purely using synthetic data from attempts to solve geometry problems.”
- Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models (2025, neutral): “…that facilitates the generation of high-quality CoT processes, which are post-processed and integrated with the original data to create the final Vision-R1-cold dataset. to progressively develop more complex reasoning processes while effectively guiding MLLMs toward enhanced reasoning capability.”