I am a Ph.D. Scholar in Indian Institute of Technology, Delhi, India working towards building resource efficient Deep Neural Network models and approaches to compress existing state of art models. My research is intended to reduce the cost of training and infering DNN models, such that complex computer vision and langauge models can be deployed to power and space constrained hardware.
Ph.D in Machine Learning • 2019-present
My research work focusses on finding novel approaches towards building deep neural networks models that can be used for inference in devices with low memory and power ie., edge devices like microcontrollers. Guide: Prof Sumeet Agarwal , Prof. Prathosh AP
M.Tech in Electronics and Communication • 2018
Being introduced to research work in the field of audio signal processing, I tinkered with tempo estimation and wrote a paper on Octave correction in Tempo estimation, which was presented in an international conference and published by IEEE xplore. Guide: Prof Subrat Bhattachrya
B.E in Electronics and Communication • 2015
During the Bachelors in engineering, I delved into various projects like Simulation of an advanced control algorithm on a Digital Signal processor that was part of the simmulaton team for lunar landing experiements. Mentor: Prof. R Muralishankar
Oshin Dutta , Ritvik Gupta, Sumeet Agarwal Paper
Oshin Dutta , Tanu Kanvar, Sumeet Agarwal, in PReMI 2023. Paper
Oshin Dutta , Ayush Srivastava, Jigyasa Gupta, Sumeet Agarwal, and Prathosh AP, in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (pp. 2745-2754), 2021. Paper Video
Oshin Dutta, in Second International Conference on Inventive Communication and Computational Technologies (ICICCT). IEEE, 2018. Paper