New Technology– category –
-
ML Interatomic Potentials: DocShare 2026 accuracy explained
DocShare Unknown Overview A study published on DocShare identified a consistent 'systematic softening' in the potential energy surfaces (PES) of universal machine-learned interatomic potentials (uMLIPs) such as M3GNet, CHGNet, and MACE-M... -
CoFe2O4: AI foundation model vs DFT site preference specs
Ouro Unknown Overview Recent analysis reveals that AI foundation models for materials science, such as CHGNet and MACE-MP-0, exhibit limitations in accurately predicting d-electron site preferences in spinel structures. Specifically, for... -
UCB-Bandit & LLM: Bayesian optimization for materials design 2026
arXiv International Overview A new paper reformulates acquisition function selection in constrained multi-objective Bayesian optimization (CMOBO) for materials design as an adaptive policy selection problem. Two novel controllers, UCB-Ba... -
Co-Based Superalloys: ML-driven 15% solvus temperature boost 2026
arXiv Unknown Overview This study introduces a physics-guided machine learning framework for the inverse design of Co-based superalloys, successfully identifying new alloy compositions that enhance the γ' solvus temperature by up to 15% ... -
UFO-MGen explained: Flow-based generative model for crystals 2026
arXiv Unknown Overview This arXiv study introduces UFO-MGen, a flow-based universal generative model that learns topological features of Wyckoff representations, achieving over 90% success rate in crystal generation. UFO-MGen demonstrate... -
MLIP training: Small-dataset strategy for 0.3 eV error in 2026
arXiv (Preprint) Unknown Overview A preprint on arXiv proposes a novel training scheme for machine-learned interatomic potentials (MLIPs) to accelerate point defect simulations. This efficient method trains MLIPs using limited DFT data, ... -
MIT 3D Generative AI: Text-to-physical object specs for 2026
Bits to Atoms (MIT CDFAM Washington DC 2026 presentation) USA Overview MIT's Alexander Htet Kyaw has unveiled a pioneering research pipeline that combines 3D generative AI, vision-language models, and robotic assembly. This system can ge... -
MLIPs: Truncated Automatic Sparse Differentiation efficiency specs
Cool Papers (Preprint/Research Highlight) Unknown Overview Calculating high-order derivatives, such as Hessians, for machine learning interatomic potentials (MLIPs) has been computationally intensive for large systems. This research prop... -
PHICS: Physics-constrained ML for composite inverse design specs
Science China Press (Research Highlight/Press Release) China Overview The Physics-constrained inverse design system (PHICS) has been developed for the efficient inverse design of multifunctional composite materials. PHICS integrates phys... -
ML Interatomic Potentials: Quantum computing data bridge explained
Cool Papers (Preprint/Research Highlight) Unknown Overview This research proposes a framework to enhance the accuracy of universal machine-learning interatomic potentials (uMLIPs) by leveraging sparse, high-fidelity data from quantum com...
