The paper identifies a phenomenon called principle-level search collapse in feedback-driven large language model (LLM) optimizers, where the search narrows to a limited set of design principles despite ongoing candidate generation and minor improvements. Using a diagnostic benchmark across molecular optimization, network sentinel-node selection, and GPU-kernel optimization, the authors find that 78.3% of runs exhibit collapse, with 55.3% showing strong support. They evaluate common remedies such as diversity injection and reference-principle interventions, noting limited and inconsistent benefits, and emphasize the need for online collapse detection and improved mechanisms for generating and maintaining alternative principles.
By Nian Ran, Tianhe Jia, Zhengyang Mu, Yue Wang, Sebastian Brulin, Markus Olhofer, Mauricio A. Alvarez, Yang Liu, Richard Allmendinger
BOOM is a new benchmark for evaluating out‑of‑distribution (OOD) molecular property predictions in machine learning. It provides chemically‑informed tests across common property prediction tasks and assesses over 150 model‑task combinations. The study shows that current models, including chemical foundation models, struggle to generalize OOD, with the best model still exhibiting three times higher error than in‑distribution predictions.
By Evan R. Antoniuk, Shehtab Zaman, Tal Ben-Nun, Peggy Li, James Diffenderfer, Busra Sahin, Obadiah Smolenski, Everett Grethel, Tim Hsu, Anna M. Hiszpanski, Kenneth Chiu, Bhavya Kailkhura, Brian Van Essen
The paper introduces MEHnet-MG, an equivariant neural network that predicts a one‑electron Hamiltonian from a single inexpensive B3LYP/def2‑SVP calculation and uses it to compute a wide range of molecular properties—energy, optical gap, dipole, quadrupole, polarizability, Mulliken charges, and Mayer bond orders—at coupled‑cluster accuracy for nine main‑group elements, including phosphorus, sulfur, and chlorine. Trained on a new CCSD(T) dataset, the model reduces property errors by factors ranging from 3.8 to 230 compared to various DFT methods while adding only ~25 ms per molecule. Importantly, by deriving properties from a predicted Hamiltonian rather than pooling atomic features, the architecture naturally incorporates correct size‑scaling, enabling accurate extrapolation to large π‑conjugated systems (up to 58 atoms) where traditional pooling‑based models fail.
By Wenhao He, Xu Chen, Noah Song, Haowei Xu, Tim S. Hindges, Bohan Li, Zihan Lin, Yu Yao, Avetik R. Harutyunyan, Fang Liu, Yao Wang, Hao Tang, Ju Li
arXiv:2607. 05736v1 Announce Type: new Abstract: Molecular property prediction often relies on isolated data modalities, where continuous 3D graph neural networks (GNNs) struggle to efficiently capture long-range topological dependencies and exact macroscopic heuristics.
By Qiwei Han, Chi Zhou, Ruobing Wang, Zheng Ma
The paper introduces a benchmark to evaluate compositional generalisation in machine learning interatomic potentials. It comprises four tasks that test models on unseen molecules, designed so that learning underlying chemical principles should enable generalisation. Experiments show that current state‑of‑the‑art models struggle, with out‑of‑distribution errors often an order of magnitude higher than in‑distribution errors, even for foundation models pre‑trained on millions of molecules.
By Amir Masoud Nourollah, Irtaza Khalid, Stefano Leoni, Steven Schockaert
FoldPipe is a lightweight Python orchestration layer that manages pre-sharded PyTorch and PyTorch Geometric molecular data. It asynchronously prefetches one shard ahead in a background thread while training on the current shard, thereby keeping the number of live shard payloads bounded relative to the total dataset size. In experiments on a SchNet energy-and-force workload with MD17 aspirin on a Tesla T4, FoldPipe achieved a mean I/O‑compute overlap of 16.33 s and a mean pass time of 76.78 s versus 83.37 s for a sequential baseline, though the overall speedup was modest and statistically inconclusive due to network variability.
By Dhiren Mukesh Khatri