Ma ZhenYao, Yue Liang, Jian-Qiao Zhu
arXiv 20 Oct 2026 · Artificial Intelligence
arXiv:2610.28562 · PDF · Extracted main text
Many scientific phenomena are best understood as outcomes of underlying optimization problems, yet automated scientific discovery has largely focused on recovering equations that describe observed input-output patterns. We introduce OptScientist, a novel automated framework for discovering compact symbolic optimization laws directly from data. OptScientist comprises (i) a model that infers symbolic optimization problems by first learning to represent data as an optimization problem and then extracting its symbolic form, and (ii) an LLM agentic system embedding the model in a discovery process that integrates theoretical priors from scientific literature, discovers candidate laws, abstracts their underlying mechanisms, and suggests new data to collect. We first evaluate OptScientist on OptFeynman, a novel benchmark of 99 physics-derived optimization problems, where OptScientist recovers ground-truth laws for over 80% of the problems and remains robust to noisy contextual information, while substantially outperforming machine-learning, symbolic, and LLM-based model-discovery baselines. On human risky choice and central-bank monetary policy, OptScientist discovers compact symbolic optimization laws that achieve held-out predictive performance competitive with leading baselines, while using few parameters and substantially fewer tokens than LLM baselines. Overall, OptScientist extends automated scientific discovery to uncovering optimization laws that generate observed phenomena, enabling automated theory building across the physical, biological, behavioral, and social sciences.
appendix boundary found by appendix_command · 40% of the source is main text. Read the extracted text to check this.
The works this paper leans on most, across its whole bibliography — not restricted to papers in our corpus. Ranked by composite intensity, which combines how often a work is mentioned, how many sections mention it, and how much of that falls in the main text rather than the appendix.
| Reference | Intensity | Mentions | Sections | Main text | |
|---|---|---|---|---|---|
| 1 | Ma, Zhenyao and Liang, Yue and Li, Dongxu (2026) Behavior Learning (BL) self | 0.941 | 6 | 4 | 83% |
| 2 | Udrescu, Silviu-Marian and Tegmark, Max (2020) AI Feynman: A physics-inspired method for symbolic regression | 0.843 | 3 | 3 | 100% |
| 3 | Plonsky, Ori and Erev, Ido and Ert, Eyal (2019) All Raw Data for CPC18: Choice Prediction Competition 2018 | 0.737 | 3 | 3 | 67% |
| 4 | Peterson, Joshua C and Bourgin, David D and Agrawal, Mayank and Reic… (2021) Using large-scale experiments and machine learning to discover theories of human decision-making | 0.737 | 3 | 2 | 100% |
| 5 | Rmus, Milena and Jagadish, Akshay Kumar and Mathony, Marvin and Ludw… (2026) Generating computational cognitive models using large language models | 0.693 | 6 | 3 | 33% |
| 6 | Champion, Kathleen and Lusch, Bethany and Kutz, J Nathan and Brunton… (2019) Data-driven discovery of coordinates and governing equations | 0.644 | 2 | 2 | 100% |
| 7 | Debreu, Gerard (1959) Theory of value: An axiomatic analysis of economic equilibrium | 0.644 | 2 | 2 | 100% |
| 8 | Lieder, Falk and Griffiths, Thomas L (2020) Resource-rational analysis: Understanding human cognition as the optimal use of limited computational resources | 0.644 | 2 | 2 | 100% |
| 9 | Zhu, Jian-Qiao and Peterson, Joshua C and Enke, Benjamin and Griffit… (2025) Capturing the complexity of human strategic decision-making with machine learning self | 0.644 | 2 | 2 | 100% |
| 10 | Murphy, Kevin (2026) Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models | 0.511 | 5 | 2 | 20% |
Showing the top 10 of 62 scored citations.