Video Research Intern at Apple, May - Aug 2023 (Summer'23)
Worked on improving the training efficiency (with coreset selection) and robutness (with augmentations) of deep learning models.
Teaching Assistant (TA) at Johns Hopkins University, August - December 2022 (Fall'22)
Teaching assistant for EN.520.344 (Fall 2022) – Introduction to Digital Signal Processing, Johns Hopkins University.
Topics discussed: Review of signals and LTI systems, Z-transform, Uniform sampling, Transform analysis of LTI systems, Filter design techniques, The discrete Fourier transform, Wavelets, Compressive sampling, Time-frequency distributions
AI Research Intern at Zippin, May - August 2022 (Summer'22)
San Francisco, CA
San Francisco, CA
I have worked on self-supervised representation learning from videos and video-based downstream tasks, including action classification and event detection.
We proposed an adaptive masking strategy (called AdaMAE) for MAEs that is end-to-end trainable. Our adaptive masking strategy samples visible tokens based on the semantic context using an auxiliary sampling network.
We demonstrated that AdaMAE samples more tokens from the high spatiotemporal information regions, thereby allowing us to mask 95% of tokens, resulting in lower memory requirements and faster pre-training.
We reported state-of-the-art results of 70.0% and 81.7% in top-1 accuracy on SSv2 and Kinetics-400 action classification datasets with a ViT-Base backbone and 800 pre-training epochs.
Teaching Assistant (TA) at Johns Hopkins University, August - December 2021 (Fall'22)
Teaching assistant for EN.520.344 (Fall 2021) – Introduction to Digital Signal Processing, Johns Hopkins University.
Topics discussed: review of signals and LTI systems, Z-transform, Uniform sampling, transform analysis of LTI systems, filter design techniques, the discrete Fourier transform, Wavelets, compressive sampling, time-frequency distributions
Research Assistant at University of Peradeniya, March - August 2020
Worked as a research assistant under the National Science Foundation Grant (Grant No: RG/2018/EA&ICT/01) “Development of a novel predictive based Smart Distribution Management System (S-DMS) to maximize the rooftop PV absorption capacity of last-mile networks” in DEEE, University of Peradeniya.