Projects

AI-AFM Assisted Structure Prediction of Protein Complexes

• Developing novel view synthesis methods for 3D NeRF reconstruction of protein structure using diffusion models.
• Created Virtual AFM:gpu-accelerated computer graphics program and accelerated data generation using GNU Parallel.
• Compiled an large dataset featuring multi-view virtual AFM images for over 550,000 voxelized 3D protein structures.

Cell Shape Detection in AFM Microscopic Images

• Implemented YOLO object detection algorithm using PyTorch for cell shape detection in AFM images.
• Incorporated intelligent vision-based navigation for traversing the AFM scanning probe.

Deep Learning for High-resolution 3D Structural Topology Optimization

• Designed PSP-U-Net architecture in Keras for performing STO.
• Leveraged multigrid-style training to train up to high-resolution of 128 × 128 × 128 3D structures.
• Achieved 5× speed up using distributed training across multiple GPUs.

Deep Learning for Structural Topology Optimization

• Designed multiple neural networks model to perform end-to-end topology optimization consistent with SIMP algorithm using 3D CNNs.
• Built and accelerated the dataset generation pipeline using GNU parallel shell and created a dataset of 60K high-resolution of 128 × 128 × 128 3D voxelized structures.