Newest papers

OptScientist: Automated Discovery of Optimization Laws

Ma ZhenYao, Yue Liang, Jian-Qiao Zhu · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28562 · cited by 0 works (OpenAlex) · open-access PDF

Indonesian

Banyak fenomena biologi yang paling baik dipahami sebagai hasil dari masalah optimalisasi yang mendasarinya, namun penemuan ilmiah otomatis sebagian besar berfokus pada pemulihan persamaan yang menggambarkan pola input-output ilmiah yang diamati. Kami memperkenalkan OptScientist, kerangka kerja otomatis baru untuk menemukan hukum optimalisasi simbolik yang kompak secara langsung dari data. OptScientist mencakup (i) sebuah model yang menurunkan mekanisme optimalisasi simbolik yang mendasarinya, dan menyarankan data baru untuk dikumpulkan dengan mempelajari pertama untuk mewakili data sebagai masalah optimalisasi dan kemudian mengekstrak bentuk simboliknya, dan (ii) sebuah sistem agentik LLM yang menanamkan model dalam proses penemuan yang mengintegrasikan prinsip dasar teoritis dari literatur ilmiah, menemukan hukum kandidat atas dasarnya, menggunakan model kebijakan dasar mereka, dan menyarankan data baru untuk

Machine translation (Meta M2M100); the English original follows.

Abstract

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.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Embedding, Benchmark (surveying), Process (computing), Optimization problem, Artificial intelligence

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange