Selected publications

For a full list of publications please refer to my 🎓 Google Scholar profile.

📜 IntelliPore: A Foundation Model for Gas Adsorption in Porous Materials (submitted)

Introduced IntelliPore, a foundation model pretrained at scale on energy images and adsorption-related data across multiple tasks and domains, enabling transfer learning across porous materials and gas adsorption properties.

📜 RetNeXt: A Pretrained Model for Transfer Learning Across the MOF Adsorption Space

Introduced a multi-task pretraining approach for learning transferable representations across the MOF adsorption space and developed RetNeXt, a 3D convolutional neural network that leverages energy images for transfer learning across gas adsorption properties.

📜 Gas adsorption meets geometric deep learning: points, set and match

Proposed a point cloud representation of porous materials and developed AIdsorb, a geometric deep learning framework for learning gas adsorption properties directly from the raw 3D structure.

📜 Gas adsorption meets deep learning: voxelizing the potential energy surface of metal-organic frameworks

Proposed a 3D representation for porous materials obtained by voxelizing the potential-energy surface, and developed RetNet, a 3D convolutional neural network for learning gas adsorption properties directly from energy images.

📜 Comparison of machine learning approaches for the identification of top-performing materials for hydrogen storage

Benchmarked machine learning approaches for efficiently screening MOFs and identifying high-performing materials for hydrogen storage.

📜 Comparison of Energy-Based Machine Learning Descriptors for Gas Adsorption

Benchmarked energy-based descriptors for machine learning prediction of gas adsorption across different gases and MOFs.