Accelerating science with AI and simulations

2026-02-12 · Source: MIT News - Machine learning · Field: Science & Research — Artificial Intelligence & Machine Learning, Physical Sciences & Chemistry, Engineering & Applied Sciences · Depth: Advanced, medium

Summary

MIT Associate Professor Rafael Gómez-Bombarelli, a newly tenured professor in materials science and engineering, believes artificial intelligence is at a second inflection point, poised to transform scientific discovery. His work combines physics-based simulations with machine learning and generative AI to discover new materials for applications in batteries, catalysts, plastics, and OLEDs. This current phase integrates language and multiple modalities into "general scientific intelligence," enabling reasoning over language, material structures, and synthesis recipes. Gómez-Bombarelli's lab, which is solely computational, focuses on how atomic composition, structure, and reactivity impact material performance, developing tools that merge deep learning with physics-based modeling. He also co-founded Lila Sciences, a company building a scientific superintelligence platform for life sciences, chemical, and materials science industries, aiming to make scientific research more seamless and productive.

Key takeaway

For AI Researchers and Research Scientists focused on materials discovery, this inflection point demands integrating language models with multi-modal scientific data. Your efforts should prioritize developing "general scientific intelligence" platforms that can reason across diverse scientific domains, moving beyond single-modality AI applications. Consider how your computational tools can create virtuous cycles with physics-based simulations to accelerate material innovation and reduce experimental overhead.

Key insights

AI is at a second inflection point, integrating language and multi-modal reasoning to accelerate scientific discovery.

Principles

• AI for science offers aspirational uses with fewer downsides.
• Physics-based simulations enhance AI algorithms with more data.
• Scaling laws apply across simulations, language, and science.

Method

Combine physics-based simulations with machine learning and generative AI to discover new materials, then merge deep learning with physics-based modeling to create computational tools.

In practice

• Apply generative AI to discover new battery materials.
• Use high-throughput simulations for material optimization.
• Develop computational tools for experimentalists to triage ideas.

Topics

• AI for Science
• Materials Discovery
• Generative AI
• Physics-based Simulations
• Deep Learning

Best for: AI Researcher, AI Scientist, Research Scientist

Related on AIssential

• 🔬Why There Is No "AlphaFold for Materials" — AI for Materials Discovery with Heather Kulik — Deep integration of domain expertise with AI and critical assessment of model outputs are crucial for success in AI for science.
• Venture Kick backs Fainite to advance physics-based simulations — Physics-aware AI accelerates complex engineering simulations, reducing costs and improving design accessibility.
• AI just discovered new physics in the fourth state of matter — AI can discover new physical laws and challenge long-standing assumptions in complex systems.
• Advancing Scientific Discovery in the Agentic AI Era — Accelerated computing and AI agents are dramatically speeding up scientific discovery beyond human processing limits.
• A Three-Layer Framework for AI in Scientific Discovery — True scientific discovery in AI hinges on qualitative reasoning for model formation, transcending mere search or execution.
• 🔬Searching the Space of All Possible Materials — Prof. Max Welling, CuspAI — AI for science is an exploding field, combining deep scientific problems with high-impact technological solutions.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by MIT News - Machine learning.