We are an independent research group focusing on applying message-passing neural networks to model potential energy surfaces and point cloud–based deep learning methods for drug discovery.
> B.Sc. Applied Chemistry — University of Tehran
> M.Sc. Physical Chemistry — Sharif University of Technology
Supervisor:
Prof. Zahra Jamshidi (Link)(Link)
E-mail: mahdi.zardoshti02@sharif.edu
CV: [Download Link]
LinkedIn: [Link]
GitHub: [Link]
> M.Sc. Biochemistry— University of Tehran
E-mail: reza.zardoshti02@ut.ac
CV: [Download Link]
LinkedIn: [Link]
GitHub: [Link]
>>> Point cloud–based deep learning for predicting drug efficacy on the VEGFR2 protein.
>>> Generative point-cloud models combined with transformer-based SMILES captioning for de novo drug design.
>>> Message-Passing Neural Networks for Small Protein Folding Pathways.
>>> Machine-learning–enhanced molecular dynamics for infrared (IR) spectral simulation.
>>> Accelerating the global search of Ag/Au clusters using message-passing neural networks and Gaussian process regression.
>>> Learning ground- and excited-state potential energy surfaces for non-adiabatic molecular dynamics.
>>> Applying transfer learning to achieve high-level (e.g., CCSD(T)-quality) potential energy surfaces.
Research Interests — Architectures
> Geometric Deep Learning
> Point Clouds Processing
> Graph Neural Networks
> Deep Tensor Neural Networks (DTNN)
> Generative Models
> Transformer Architecture
> Long Short-Term Memory (LSTM)
Home | Research | Contact
Deep Learning for Computational Chemistry and Drug Design
Zardosht Research Group
Tehran/Iran
