Research

Publications

• [Re] Training Binary Neural Networks using the Bayesian Learning Rule
  Prateek Garg, Lakshya Singhal, Ashish Sardana
  ReScience C Journal, ML Reproducibility Challenge
• Short-term traffic prediction under non-recurrent congestion: A machine learning approach for traffic management during disaster
  Makoto Chikaraishi, Prateek Garg, Varun Varghese, Kazuki Yoshizoe, Junji Urata, Yasuhiro Shiomi, Ryuki Watanabe
  2019

Academic Projects

• Traffic Adaptability (B.Tech Thesis) with Prof. Rijurekha Sen: meta-learning (MAML) and continual learning so traffic models keep adapting after deployment.
• Spine Classification & Segmentation with Prof. Prathosh AP: U-Net segmentation and transfer learning on X-ray images.
• AI Bots with Prof. Mausam: simulated annealing for gene mapping, an alpha-beta bot for the Game of Cannon, and SAT-based graph mapping.
• Encrypted Chat Application with Prof. Aaditeshwar Seth: end-to-end encrypted messaging that the server cannot decrypt.

Patents

Filed from my work at Honda Innovation Lab, Tokyo.

• Image recording device and medium
  Patent Application US20260025484A1
• Driver monitoring apparatus and medium
  Patent Application US20260024358A1
• Risk presentation apparatus and risk presentation method
  Patent Application US20260021824A1
• Driver-assistance apparatus, driver-assistance method, and storage medium
  Patent Application US20260024353A1
• Drive recorder and control method thereof
  Patent Application US20260015007A1
• Driver-assistance apparatus, driver-assistance method, and storage medium
  Patent Application US20260018053A1

Past Supervisors

• Jovin D'sa
  Principal Engineer, Honda Research Institute, Tokyo
• Prof. Martin Mueller
  Canada CIFAR AI Chair at Amii, University of Alberta
• Dr. Kazuki Yoshizoe
  Professor, Kyushu University, Japan