{"total":14,"items":[{"citing_arxiv_id":"2607.01408","ref_index":9,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Spin-Weighted Spherical Harmonics Enable Complete and Scalable $\\mathrm{E}(3)$-Equivariant Networks","primary_cat":"cs.LG","submitted_at":"2026-07-01T19:13:44+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"SpinGTP uses Spin-Weighted Spherical Harmonics to complete the Gaunt Tensor Product, achieving full E(3)-equivariance with GTP-like scaling and better handling of chiral and parity-odd cases.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.29584","ref_index":4,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Geometric Algebra Meets Cartesian Tensors: Higher-Order Equivariance for Interatomic Potentials","primary_cat":"physics.chem-ph","submitted_at":"2026-06-28T19:57:50+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"CliffordSTF couples Clifford multivectors to rank-2 and rank-3 symmetric-traceless tensor tracks through bilinear cross-track contractions, lifting force cosine similarity from 0.055 to 0.551 on rMD17 while outperforming CG-free baselines.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.09480","ref_index":12,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Loss-Guided Adaptive Scale Refinement for Molecular Force Prediction","primary_cat":"cs.LG","submitted_at":"2026-06-08T13:39:02+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Loss-guided adaptive scale refinement on NaCl aqueous system reduces overall force MAE from 399.65 to 381.23 by discovering intermediate scales from initial anchors.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.02785","ref_index":12,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"QUIVER: Quantum-Informed Views for Enhanced Representations in Large ML Models","primary_cat":"cs.LG","submitted_at":"2026-06-01T18:50:28+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"QUIVER augments classical ML models with a quantum Fisher view from VQCs to improve performance on QM9 molecular properties and JetClass jet flavor prediction.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.02419","ref_index":195,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"DPA4: Pushing the Accuracy-Cost Frontier of Interatomic Potentials with EMFA SO(2) Convolution","primary_cat":"physics.chem-ph","submitted_at":"2026-06-01T15:59:10+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"DPA4 is a new SE(3)-equivariant interatomic potential with EMFA SO(2) convolution that sets new accuracy-cost records on Matbench Discovery and SPICE benchmarks using fewer parameters than prior models.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.29329","ref_index":18,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Mixing Vector Model for Copolymer Inference via Mixed Integer Linear Programming","primary_cat":"q-bio.QM","submitted_at":"2026-05-28T04:05:30+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Extends the mol-infer MILP framework to copolymers via a mixing vector representation, trains ML predictors achieving R^2 >0.7 on 9/10 datasets, and shows tractable multi-monomer inverse design with external consistency checks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.07567","ref_index":21,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"SurfDesign: Effective Protein Design on Molecular Surfaces","primary_cat":"q-bio.BM","submitted_at":"2026-05-25T19:53:02+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SurfDesign introduces surface-conditioned protein design via manifold modeling and equivariant message passing on surfaces integrated with pretrained language models, outperforming prior methods on binder and enzyme design benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.16891","ref_index":10,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Tensor Channel Equivariant Graph Neural Networks for Molecular Polarizability Prediction","primary_cat":"cs.LG","submitted_at":"2026-05-16T09:07:07+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"A tensor-channel equivariant GNN based on PaiNN propagates symmetric rank-2 tensor features during message passing and achieves lower full-tensor and anisotropic error than readout-only and MACE baselines on QM7-X geometries.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.14154","ref_index":22,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"TSAgent: An Agentic Workflow for Autonomous Transition State Search","primary_cat":"physics.chem-ph","submitted_at":"2026-05-13T22:08:24+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"TSAgent automates transition state searches at DFT accuracy via an agentic loop, reaching 83% success on 100 OC20NEB examples and 70% on 10 held-out cases versus 73% for human experts.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.08988","ref_index":16,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Benchmarking Compositional Generalisation for Machine Learning Interatomic Potentials","primary_cat":"cs.LG","submitted_at":"2026-05-09T15:12:59+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"A new benchmark finds that state-of-the-art ML interatomic potentials struggle with compositional generalization, producing errors an order of magnitude higher on unseen molecular combinations than on training-like cases.","context_count":1,"top_context_role":"baseline","top_context_polarity":"baseline","context_text":"to the symbolic recombination of learned functional patterns, probing the model's capacity to handle hetero-functionalisation. 4 Evaluation 4.1 Experimental set-up We use our benchmark to evaluate a diverse set of state-of-the-art MLIPs. These include invariant message-passing networks (SchNet [40]), models incorporating explicit geometric features (GemNet [18, 17], DimeNet++ [16]), rotationally equivariant architectures (PAINN [39], NequIP [6], eSCN [32]), MACE [ 4]), and equivariant transformers (EquiFormerV2 [ 25]). We also evaluate UMA (Small) [42] and MACE-MP-0 (Small, Medium, Large) [5], as representative foundation models. Evaluating foundation models against our PBE labels presents distinct challenges. Total energies are"},{"citing_arxiv_id":"2604.23134","ref_index":9,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"h-MINT: Modeling Pocket-Ligand Binding with Hierarchical Molecular Interaction Network","primary_cat":"cs.LG","submitted_at":"2026-04-25T04:25:50+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"h-MINT improves ligand-protein binding affinity prediction by 2-4% and virtual screening metrics by 1-3% via overlapping fragment tokenization and hierarchical modeling.","context_count":1,"top_context_role":"method","top_context_polarity":"use_method","context_text":"of at to as, ei,j is the edge type embedding of (fi, fj)8, and Wα is to project the scores into scalars. Through Softmax in Eq. 8,a s is able to aggregate messages froma t ∈knn(f j, as). Thetoken-level cross attentionis defined through aggregating all atom-level edges expanded from the token-level edge. Score:S i,j = 1 |knn(fi, fj)| X (as,at)∈knn(fi,fj) Mi,j[as, at],(9) Attention Weight:β i,j =Softmax fj ∈KNN(fi)(Si,jWβ),(10) where Wβ projects the scores into scalars. Basically, Si,j aggregates all atom-level edges expanded from(f i, fj). Then we have the following message passing and embedding update: mi,j[as] = X at∈knn(fj ,as) αi,j[as, at]Vl[at],(11) mi[as] = X fj ∈KNN(fi) βi,jMLP(mi,j[as]),(12) H l[as]←H l−1[as] +ScatterMean(m i[as],T f2a)."},{"citing_arxiv_id":"2604.16586","ref_index":96,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"A Systematic Survey and Benchmark of Deep Learning for Molecular Property Prediction in the Foundation Model Era","primary_cat":"cs.LG","submitted_at":"2026-04-17T15:16:33+00:00","verdict":"ACCEPT","verdict_confidence":"MODERATE","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A systematic survey and benchmark of four deep learning paradigms for molecular property prediction that organizes the field, critiques current data practices, and outlines three future directions.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"[94] Yi Liu, Limei Wang, Meng Liu, Yuchao Lin, Xuan Zhang, Bora Oztekin, and Shuiwang Ji. Spherical message passing for 3d molecular graphs. InInternational Conference on Learning Representations, 2022. [95] Shuqi Lu, Zhifeng Gao, Di He, Linfeng Zhang, and Guolin Ke. Highly accurate quantum chemical property prediction with uni-mol+.arXiv preprint arXiv:2303.16982, 2023. [96] Johannes Gasteiger, Shankari Giri, Johannes T Margraf, and Stephan Günnemann. Fast and uncertainty-aware directional message passing for non-equilibrium molecules.arXiv preprint arXiv:2011.14115, 2020. [97] Nathaniel Thomas, Tess Smidt, Steven Kearnes, Lusann Yang, Li Li, Kai Kohlhoff, and Patrick Riley. Tensor field networks: Rotation-and translation-equivariant neural networks for 3d point clouds."},{"citing_arxiv_id":"2604.09320","ref_index":4,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Transferable FB-GNN-MBE Framework for Potential Energy Surfaces: Data-Adaptive Transfer Learning in Deep Learned Many-Body Expansion Theory","primary_cat":"physics.chem-ph","submitted_at":"2026-04-10T13:43:19+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"FB-GNN-MBE integrates fragment-based graph neural networks into many-body expansion to predict two- and three-body energies for water, phenol, and mixture systems at chemical accuracy, with a teacher-student protocol enabling transfer to new cluster sizes without full retraining.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2603.26499","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"AIRA_2: Overcoming Bottlenecks in AI Research Agents","primary_cat":"cs.AI","submitted_at":"2026-03-27T15:02:43+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":6.0,"formal_verification":"none","one_line_summary":"AIRA₂ improves AI research agents via asynchronous multi-GPU workers, hidden consistent evaluation, and interactive ReAct agents, reaching 81.5-83.1% percentile rank on MLE-bench-30 and exceeding human SOTA on 6 of 20 AIRS-Bench tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}