mlsanityData to 13 Oct 2026

TRACE: Transition-Aware Residual Control for Multi-Objective Materials Discovery

arXiv 2608.23631 ↗First posted 23 Aug 2026cs.AI cond-mat.mtrl-scicited by 0

On the map: LLM agents and tool use → Medical and scientific domain agents

Abstract (arXiv)

Multi-objective materials discovery with LLM agents is often limited not only by how many candidates can be proposed, but by how effectively each costly property evaluation informs the next search step. Existing agents mainly store evaluated candidates and their scores, so they know which materials succeeded but not which executable edits caused useful property changes. This makes local refinement difficult when objectives compete and an edit that improves one property may damage another. We propose TRACE, a transition-aware residual control framework that treats evaluated edits as the basic unit of feedback. TRACE records each local refinement as a parent-edit-child transition with observed property deltas, aggregates transition evidence to estimate reusable edit effects, and ranks future edits by their predicted ability to reduce the current candidate's remaining constraint violations while avoiding damage to already satisfied objectives. In a controlled same-backbone comparison, TRACE improves over LLEMA, the state-of-the-art LLM-agent baseline, raising macro-average hit rate from 18.13% to 25.96%.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Placed in Medical and scientific domain agents, one of 16 directions in LLM agents and tool use. Builds on 19 of its 38 references that are in the map. First posted on arXiv 23 Aug 2026.

Builds on

19 of 38 references.

References · 19

  1. 2019When to Trust Your Model: Model-Based Policy Optimizationcited by 756
  2. 2018Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Modelscited by 754
  3. 2023ChemCrow: Augmenting large-language models with chemistry toolscited by 622
  4. 2021Crystal Diffusion Variational Autoencoder for Periodic Material Generationcited by 270
  5. 2020Differentiable Expected Hypervolume Improvement for Parallel Multi-Objective Bayesian Optimizationcited by 177
  6. 2019Symmetry-adapted generation of 3d point sets for the targeted discovery of moleculescited by 171
  7. 2022Sample Efficiency Matters: A Benchmark for Practical Molecular Optimizationcited by 151
  8. 2023Crystal Structure Prediction by Joint Equivariant Diffusioncited by 145
  9. 2021MARS: Markov Molecular Sampling for Multi-objective Drug Discoverycited by 125
  10. 2020Sample-Efficient Optimization in the Latent Space of Deep Generative Models via Weighted Retrainingcited by 122
  11. 2018Learning Multimodal Graph-to-Graph Translation for Molecular Optimizationcited by 88
  12. 2024Efficient Evolutionary Search Over Chemical Space with Large Language Modelscited by 59
7 more
  1. 2024LLMatDesign: Autonomous Materials Discovery with Large Language Modelscited by 42
  2. 2020A Deep Generative Model for Molecule Optimization via One Fragment Modificationcited by 36
  3. 2024ChemReasoner: Heuristic Search over a Large Language Model's Knowledge Space using Quantum-Chemical Feedbackcited by 32
  4. 2023Large Language Models for Intent-Driven Session Recommendationscited by 30
  5. 2023Bayesian Federated Learning: A Surveycited by 22
  6. 2025Robust Machine Unlearning for Quantized Neural Networks via Adaptive Gradient Reweighting with Similar Labelscited by 12
  7. 2025LLEMA: Evolutionary Search with LLMs for Multi-Objective Materials Discoverycited by 12

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026MADE: Benchmark Environments for Closed-Loop Materials Discovery
  2. 2026Towards Agentic Intelligence for Materials Science
  3. 2025Hypothesis Generation for Materials Discovery and Design Using Goal-Driven and Constraint-Guided LLM Agents
  4. 2025Aligning Reasoning LLMs for Materials Discovery with Physics-aware Rejection Sampling
  5. 2026Self-Revising Discovery Systems for Science: A Categorical Framework for Agentic Artificial Intelligence
  6. 2026ARIA: A Causal-Aware Framework for Rescuing LLM Reasoning in Trustworthy Materials Discovery
  7. 2026CompMat-Bench: Benchmarking AI Agents for Computational Materials Science
  8. 2025Operationalizing Serendipity: Multi-Agent AI Workflows for Enhanced Materials Characterization with Theory-in-the-Loop