SAGE-Net: Semantics-Augmented Geometric Encoder for Material Property Prediction

Dual-Level Atomic and Coordination Geometry Learning for Crystal Property Prediction Using Graph Neural Networks

The proposed Coordination Polyhedron Graph Network (CPGN) is a multi-scale GNN that jointly learns atomic, bond, and coordination-polyhedron representations and outperforms existing state-of-the-art GNN models.

GDGraph: Geometry-Enhanced Dual-View Graph for Molecular Representation Learning

Learning effective molecular representations is crucial for accurate property prediction in AI-aided drug discovery. However, most existing molecular pre-training methods are still primarily based on 2D topological graphs, limiting their ability to exploit 3D geometric information. Moreover, methods that do incorporate 3D geometry often do not distinguish between the roles of atom-centered and bond-centered representations. To address these limitations, we propose GDGraph, a geometryenhanced dual-view framework for molecular representation learning. GDGraph models molecular geometry from two complementary structural perspectives: an atom view for capturing global spatial dependencies and a bond view for modeling local geometric patterns. To support this dual-view design, we introduce a multi-scale geometric feature encoding scheme and a view-specific geometry-aware learning strategy, enabling each view to focus on the geometric dependencies it is best suited to capture. Extensive experiments demonstrate that GDGraph achieves strong and stable performance on molecular property prediction benchmarks, and effectively predicts geometrysensitive quantum chemical properties on the QM9 dataset.

MatBind: A Shared Embedding Space for Multimodal Materials Characterization

Fully characterizing a crystalline material requires integrating heterogeneous data sources -- atomic structures, diffraction patterns, electronic density of states, and natural language -- each of which captures a different facet of the same physical object. In practice, however, these modalities are stored and analyzed in isolation, making it difficult to relate or query materials across representational boundaries. We present MatBind, a contrastive learning framework that aligns four materials modalities -- crystal structure, powder X-ray diffraction (pXRD) simulated from structures, density of states (DOS), and text -- into a unified embedding space using crystal structure as the central physical anchor. The framework induces alignment between modalities never explicitly paired during training, enabling emergent zero-shot cross-modal retrieval as a direct consequence of the shared representation. The learned embedding space organizes materials according to physically meaningful properties without explicit supervision, and retrieval performance improves systematically when modalities are combined at query time. These results demonstrate that treating heterogeneous materials data as complementary projections of a single physical reality, rather than as isolated data sources, is not a practical choice but is consistent with the underlying physics.

UniPocket: Physics-Aware Geometric Graph Learning with Manifold Completeness for Ligand-Specific Binding Site Prediction

Model Agnostic Graph Prompt Learning for Crystal Property Prediction

Graph Neural Networks have emerged as a powerful tool for the fast and accurate prediction of various crystal properties. These models often encode domain-specific knowledge into their graph encoding modules, which increases their parameter size and makes their performance heavily dependent on domain expertise. Added to this, explicitly incorporating all chemical and structural features, that might influence a specific crystal property into the GNN encoder, is a challenging task. In this work, we propose a soft prompt learning framework that captures latent features essential for property prediction, which are not explicitly provided to the GNN. We introduce a novel multilevel graph prompt learning framework comprising both node-level and graph-level soft prompts. At the node level, we capture the local chemical semantics of different atom types, while at the graph level, we encode the global structural symmetry of the crystal graph. Our proposed prompt learning framework is lightweight and seamlessly integrates with any existing GNN encoder. Extensive experiments on popular benchmark datasets show that incorporating prompt learning significantly improves (3\% - 15\%) the performance of state-of-the-art GNN models in crystal property prediction tasks. Furthermore, the learned soft prompts enable cross-property knowledge transfer, enhancing prediction performance for properties with limited training data. Code is available at https://github.com/shrimonmuke0202/Prompt.git

Physics informed inverse discovery of potential topological materials via descriptor-graph ensemble learning

Topological materials (TMs) constitute a fundamentally new class of quantum matter, hosting symmetry protected electronic states with robust transport and spin–momentum locking that are central to next generation electronic, spintronic, and quantum technologies. However, their identification is conventionally driven by first principles electronic structure calculations and explicit evaluation of topological invariants, which limits scalability across large chemical spaces. Here, we present a physics informed machine learning (ML) framework for discovering topologically non-trivial materials without repeated evaluation of wavefunction based topological invariants during candidate screening, integrating electronic structure–derived proxy descriptors, symmetry-resolved topological databases, tabular ML models, and crystal graph neural networks (GNNs) into a unified prediction and inverse discovery pipeline. Using symmetry indicator derived Z2 invariants (ν0, ν1, ν2, ν3) as supervised labels, we curate a database of 9,142 crystalline materials spanning multiple dimensionalities and construct a ML dataset comprising 6,436 two- and three-dimensional materials and construct physically motivated descriptors capturing band gaps, bandwidths, Fermi level crossings, and dispersion characteristics alongside composition and structure-based representations. Tabular models achieve ROC–AUC values > 0.8, while GNNs trained solely on atomic connectivity achieve > 0.65, suggesting that crystal structure contains useful but incomplete information related to topological propensity. In an inverse discovery setting, the models assign probabilistic scores and enable consensus ranking that filters approximately 45,000 candidates to about 1,200 candidate materials for subsequent validation. We shortlist 100 top-ranked materials for future validation using first principles electronic structure calculations and topological studies. Overall, this work establishes a scalable and physics guided candidate prioritization framework for inverse discovery of topological materials beyond brute force electronic structure calculations.