Characterizing the Discrete Geometry of ReLU Networks
Published in International Conference on Learning Representations (ICLR), 2026
Analyzes the polyhedral structure induced by ReLU networks to better understand their discrete geometry.
Publications
Conference Papers
Mining Large-Scale Knowledge Graphs for Chemical Reaction Fingerprints
Published in 2023 IEEE International Conference on Big Data, 2023
Methods for extracting chemical reaction fingerprints from large-scale knowledge graphs for downstream modeling.
Heterogeneous Graph Sparsification for Efficient Representation Learning
Published in IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2022
Sparsification techniques for heterogeneous graphs to improve the scalability of representation learning.
Journal Articles
Explaining Graph Neural Networks with Mixed-Integer Programming
Published in Neurocomputing, 2025
A mixed-integer programming framework for generating faithful explanations of graph neural network predictions.
Fragment-based deep molecular generation using hierarchical chemical graph representation and multi-resolution graph variational autoencoder
Published in Molecular Informatics, 2023
A fragment-based deep generative model for molecular design using a hierarchical chemical graph representation.
OctSurf: Efficient hierarchical voxel-based molecular surface representation for protein-ligand affinity prediction
Published in Journal of Molecular Graphics and Modelling, 2021
A novel 3D deep learning approach for protein-ligand binding affinity prediction using octree-based volumetric representation to efficiently characterize molecular surfaces.