Tohoku-led team uses hybrid sampling to build ML potentials for hydrogen adsorption in MOF-303
CatAstro_Piyush · x · 2026-10-07
A collaboration led by Tohoku University's AIMR with Osaka University and AIST published a new paper in Physical Review Materials on a hybrid sampling approach for machine-learning potentials (MLPs) applied to gas adsorption, using hydrogen adsorption in the MOF-303 metal-organic framework as the case study.
Key points:
- The method combines sampling strategies to generate training data covering the configurations that matter for adsorption, improving the reliability of MLP predictions of adsorption energies and sites.
- Benchmarked on H2 adsorption in MOF-303, the MLP reproduces DFT-level accuracy.
- Implication: a scalable computational tool for screening porous materials for hydrogen storage — an applied AI-for-Science result. Open access (CC BY 4.0).
More from Research
- NVIDIA's LoGRA cuts LLM RL training memory by up to 45.7%, enables 27B RL on a single 8-GPU node — nvidia · 2026-10-07
- Microsoft's PrisMem evolves agent memory per-capability, beats baselines by 10.5 points on BEAM-1M — microsoft · 2026-10-07
- Google's SEER adds self-evolving event reasoning to time-series forecasting, beats SOTA on six benchmarks — google · 2026-10-07
- No-LLM-agent SRE diagnosis pipeline passes 80/105 cases across 21 fault scenarios in 14.6s median — tianyin_xu · 2026-10-07
- AI turns OpenAI's 199-page zeta proof into a 2-minute 3D animation claiming first zero-free strip past 7/8 — imjustnewatai · 2026-10-07
- OpenAI Reportedly Makes Serious Breakdown Toward Riemann Hypothesis — ipeirotis · 2026-10-07