Principia uses AI to discover scientific principles during testing
Principia: Test-Time Hypotheses to Principle Construction for Scientific Discovery
Machine LearningArtificial Intelligence
Summary
Finding the rules that explain scientific observations usually takes a lot of expert knowledge and time. The authors created Principia, a system using large language models to develop and improve scientific rules while experiments happen. Principia forms guesses about causes, checks them with new data, and updates its ideas continuously. This makes it easier and faster to find helpful scientific principles in various complex fields like molecular biology and protein design. The system’s principles match well with real scientific knowledge and help improve prediction tasks.
What this means in practice
- •For biotech engineers: Automate the formulation and refinement of hypotheses to guide molecular design and protein engineering experiments.
- •For drug discovery teams: Enhance molecular property predictions by dynamically updating scientific principles during experimental testing.
Authors
Dongki Kim, Sangwoo Park, Jinheon Baek, Soyeong Jeong, Sung Ju Hwang
Abstract
Scientific discovery requires identifying the scientific principles underlying observed phenomena to guide further investigation. As these empirical principles are often unknown in complex discovery processes, they are constructed via a hypotheses-to-principle pipeline to obtain recurring causalities from progressively accumulated observations during test-time discovery. However, constructing this knowledge requires substantial domain expertise, time, and effort, making this translation stage a bottleneck for scientific discovery. To address this limitation, we propose Principia, an LLM-based framework that constructs and continually refines scientific principles at test time. Drawing on how human researchers develop principles, Principia forms comparative causal hypotheses from related inputs and their observed outcomes, then synthesizes recurring relationships across these hypotheses into inductive principles. By applying the constructed hypotheses and inductive principles to the task, it obtains new observations, revises the hypotheses used based on these new observations, and reconstructs principles from the updated hypotheses, allowing scientific knowledge to continuously evolve throughout test-time discovery. Across benchmarks spanning molecular optimization, protein engineering, gene perturbation, and molecular property prediction, Principia outperforms baselines relying on internal knowledge, external sources, or experience summaries, while generating principles that are plausible and consistent with the scientific literature, demonstrating its effectiveness in constructing and refining task-specific hypotheses and principles at test time for scientific discovery. Our code is available at https://github.com/DongkiKim95/Principia.