With a passion for Computer Science, I have both the engineering skills and research background necessary to solve difficult problems in the field of Machine Learning and Computer Vision. Currently, I am working as a Research Assistant in LCI Lab at CS Department, UBC. My current research is focussed on the problems that lie at the interface of Computer Vision and Natural Language, such as video captioning, temporal localization in videos. I am always happy to connect and collaborate with the people from academia and the industry with similar interests.
Enhanced Visual Dialog
Formulated the problem as a cooperative game between two bots that communicate over a given image and made the model end-to-end differentiable using gumbel-softmax sampler.
Learning TIC using multi-armed bandits
Proposed and implemented multi bandits framework for learning topic-aware influence probabilities in social network graph.
Improved run-time by implementing online Expectation Maximization algorithm to update the probabilities.
Deep Generative Models: VAE and GAN
Survey paper discussing variational auto-encoders and generative adversarial networks. We highlight the solutions to the major challenges faced in their research. Additionally, we discuss recent research which tries to unify these models and suggest possible future research directions.
© 2023 By Mohit Bajaj. Proudly created with Wix.com