{"total":11,"items":[{"citing_arxiv_id":"2606.25742","ref_index":15,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Barocaloric phase transformation from data efficient fine-tuning of machine learned interatomic potentials","primary_cat":"cond-mat.mtrl-sci","submitted_at":"2026-06-24T12:11:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Fine-tuning the MACE-MPA-0 foundation model on 5-10 60-atom DFT configurations reproduces the barocaloric phase transformation in ammonium sulfate, while training from scratch fails at these sizes.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.11038","ref_index":28,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Synthetic pre-training of graph-network models for predicting solid-state NMR parameters","primary_cat":"cond-mat.mtrl-sci","submitted_at":"2026-06-09T16:09:01+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Synthetic pre-training on ML-generated tensor data followed by fine-tuning on ground-truth calculations improves data efficiency for graph models of solid-state NMR parameters when the pre-training and fine-tuning domains match.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.07327","ref_index":235,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Six Open Questions in Machine-Learned Interatomic Potential Foundation Models","primary_cat":"cond-mat.mtrl-sci","submitted_at":"2026-06-05T14:45:06+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":2.0,"formal_verification":"none","one_line_summary":"This perspective article develops a definition of foundational MLIPs and poses six open questions that the authors believe will define future research in machine-learned interatomic potentials.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.05127","ref_index":12,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Non-covalent Interactions at cm$^{-1}$ Accuracy: Data Efficient Physics-Informed Distillation for Machine Learning Interatomic Potentials","primary_cat":"physics.chem-ph","submitted_at":"2026-06-03T17:31:32+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Physics-informed distillation from a universal MLIP plus limited CCSD(T) fine-tuning yields cm^{-1} accurate potentials for non-covalent interactions, with teacher choice strongly affecting accuracy on some systems.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.03232","ref_index":70,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"GFFMERGE: Efficient Merging of Graph Neural Force Fields and Beyond","primary_cat":"cs.LG","submitted_at":"2026-06-02T06:48:34+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"GFFMERGE formulates GNN force field merging as a convex embedding-alignment problem with an analytical solution, recovering near joint-training performance on MD17, MD22, LiPS20 and other benchmarks while delivering 5-27x speedups.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.20581","ref_index":31,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"TriForces: Augmenting Atomistic GNNs for Transferable Representations","primary_cat":"cs.LG","submitted_at":"2026-05-20T00:38:43+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"TriForces adds a model-agnostic three-stream architecture plus self-supervised objectives to atomistic GNNs, improving transfer performance on MatBench, QM9, and limited-data OMat24 without DFT labels.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.27879","ref_index":27,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Generation of magnetic metal-organic frameworks","primary_cat":"cond-mat.mtrl-sci","submitted_at":"2026-04-30T13:56:52+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Fine-tuning CHGNet on OMDB data and performing site substitution on QMOF prototypes yields novel highly magnetic MOFs.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.27351","ref_index":82,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Heterogeneous Scientific Foundation Model Collaboration","primary_cat":"cs.AI","submitted_at":"2026-04-30T03:02:27+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Eywa enables language-based agentic AI systems to collaborate with specialized scientific foundation models for improved performance on structured data tasks.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"data, TabPFN [54] and its successors [78, 79] leverage in-context learning to solve prediction tasks and outperform tuned tree-based ensembles. The broader ecosystem of domain-specific foundation models further extends to even more specialized scientific representations. In materials science, universal machine-learned interatomic potentials such as GNoME [80], MACE-MP-0 [81], and CHGNet [82] are trained on broad crystallographic databases and generalize across the periodic table. In weather and climate, GraphCast [24], Pangu-Weather [83], Aurora [84], and GenCast [85] have matched or surpassed operational numerical weather prediction systems for medium-range forecasting [19]. In life sciences, AlphaFold [86] and the ESM family [87, 88] have transformed protein structure prediction and design through large-scale pretraining on"},{"citing_arxiv_id":"2604.25756","ref_index":40,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Predicting challenging phase transitions with Bayesian active learning","primary_cat":"cond-mat.mtrl-sci","submitted_at":"2026-04-28T15:19:43+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Bayesian active learning with SSCHA predicts phase transitions in materials like CsPbI3 using only 50-256 first-principles calculations.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.21494","ref_index":46,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Data-Driven Thermal and Mechanical Modeling of Defective Covalent Organic Frameworks","primary_cat":"cond-mat.mtrl-sci","submitted_at":"2026-04-23T09:54:53+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"QCOF ML potentials tuned on COF data outperform general MACE models for defective systems and reveal higher thermal defect sensitivity in CTF-1 versus COF-LZU1 with nearly invariant low-strain mechanics.","context_count":1,"top_context_role":"dataset","top_context_polarity":"use_dataset","context_text":"forces from a diverse dataset comprising over 36,000 conformations of 23 distinct carbon nitride nanosheets, computed using DFT at the PBE+D3 level of theory. The dataset consists of periodic 2D structures containing approximately 100 atoms each. In addition to equilibrium configurations, it includes structures subjected to large tensile strains up to fracture. Further details can be found in the original database publication, see Ref. [46]. Within the MACE architecture, three key hyperparameters were varied: the atomic descriptor dimensionality D (ranging from 32 to 256), the inclusion of angular equivariance E (E=0 for invariance, E=1 for equivariance), and the interaction cutoff radius rc (ranging from 3 to 7 Å). The resulting models are labelled according to these hyperparameters, namely, D−E−r c."},{"citing_arxiv_id":"2603.10992","ref_index":21,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"A Tutorial Review of Bayesian Optimization with Gaussian Processes to Accelerate Stationary Point Searches","primary_cat":"stat.ML","submitted_at":"2026-03-11T17:20:28+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"Bayesian optimization with Gaussian processes unifies minimization, single-point saddle searches, and double-ended path searches on potential energy surfaces through a shared six-step surrogate loop using derivative observations and inverse-distance kernels.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}