Tag
MIT researchers introduced a framework called CrysVCD that enhances the chemical stability of AI-generated materials by enforcing valence rules, making them more suitable for real-world applications like computer chips and rockets.
The paper proposes TRACE, a transition-aware residual control framework for multi-objective materials discovery using LLM agents, which improves hit rates over the state-of-the-art baseline.
This article explores whether current AI models could drive breakthrough scientific discoveries, such as new materials, by accelerating hypothesis generation and research, drawing parallels to the LK-99 superconductor episode.
Discovered Materials, a YC P26 startup, launches AI agents for discovering new materials, showcasing a benchmark where frontier LLMs find stable new materials but struggle to propose plausible synthesis recipes.
Advaith Sridhar introduces Discovered Materials, a startup building AI scientists to discover new semiconductor materials, releasing hundreds of discoveries and a benchmark.
The LEAP framework integrates a domain-specialized large language model with active learning to efficiently prioritize precursor additives for perovskite solar cells, achieving improved power conversion efficiencies in experimental validation.
CrystalReasoner is an LLM framework that generates crystal structures from natural language by using physical priors as thinking tokens and reinforcement learning to ensure validity, stability, and property-conditioned generation.