mlsanityData to 7 Oct 2026

On The Finetuning of MLIPs Through the Lens of Iterated Maps With BPTT

arXiv 2512.01067 ↗First posted 30 Nov 2025cond-mat.mtrl-sci cs.AI cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Accurate structural relaxation is critical for advanced materials design. Traditional approaches built on physics-derived first-principles calculations are computationally expensive, motivating the creation of machine-learning interatomic potentials (MLIPs), which strive to faithfully reproduce first-principles computed forces. We propose a fine-tuning method to be used on a pretrained MLIP in which we create a fully-differentiable end-to-end simulation loop that optimizes the predicted final structures directly. Trajectories are unrolled and gradients are tracked through the entire relaxation. We show that this method consistently improves performance across all evaluated pretrained models; resulting in an average of roughly 32% reduction in prediction error. Interestingly, we show the process is robust to substantial variation in the relaxation setup, achieving negligibly different results across varied hyperparameter and procedural modifications.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 26 of its 69 references that are in the map. First posted on arXiv 30 Nov 2025.

Builds on

26 of 69 references.

References · 26

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  3. 2017Proximal Policy Optimization Algorithmscited by 14,441
  4. 2017Deep reinforcement learning from human preferencescited by 3,846
  5. 2019Fine-Tuning Language Models from Human Preferencescited by 2,124
  6. 2019Deep Equilibrium Modelscited by 603
  7. 2022MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fieldscited by 467
  8. 2022Learning Local Equivariant Representations for Large-Scale Atomistic Dynamicscited by 254
  9. 2020The Open Catalyst 2020 (OC20) Dataset and Community Challengescited by 246
  10. 2018Explicit Inductive Bias for Transfer Learning with Convolutional Networkscited by 226
  11. 2024Secrets of RLHF in Large Language Models Part II: Reward Modelingcited by 170
  12. 2021Stabilizing Equilibrium Models by Jacobian Regularizationcited by 71
14 more
  1. 2025Command A: An Enterprise-Ready Large Language Modelcited by 57
  2. 2018Amortized Inference Regularizationcited by 30
  3. 2025A practical guide to machine learning interatomic potentials -- Status and futurecited by 24
  4. 2024Does RLHF Scale? Exploring the Impacts From Data, Model, and Methodcited by 22
  5. 2023Wide Neural Networks as Gaussian Processes: Lessons from Deep Equilibrium Modelscited by 14
  6. 2022Deep equilibrium networks are sensitive to initialization statisticscited by 8
  7. 2024Positive concave deep equilibrium modelscited by 7
  8. 2025Deep Equilibrium models for Poisson Imaging Inverse problems via Mirror Descentcited by 7
  9. 2024Infusing Self-Consistency into Density Functional Theory Hamiltonian Prediction via Deep Equilibrium Modelscited by 6
  10. 2025Transformers Discover Molecular Structure Without Graph Priorscited by 6
  11. 2022Improving Stability of Fine-Tuning Pretrained Language Models via Component-Wise Gradient Norm Clippingcited by 5
  12. 2025Learning Inter-Atomic Potentials without Explicit Equivariancecited by 5
  13. 2024Understanding Representation of Deep Equilibrium Models from Neural Collapse Perspectivecited by 3
  14. 2025ADAPT: Lightweight, Long-Range Machine Learning Force Fields Without Graphscited by 1

Built on by

Not cited yet.