Jathushan Rajasegaran

I am a Ph.D. student at BAIR advised by Prof. Jitendra Malik. I am also visiting researcher at Meta AI working with with Dr. Christoph Feichtenhofer. I am broadly interested in Computer Vision and Deep Learning, with a focus on developing models for visual understanding in images and videos. Before coming to Berkeley, I was working with Prof. Salman Khan at Inception Institute. I completed my undergraduate study at University of Moratuwa, with a major in Electronic and Telecommunication Engineering. My Bachelor's Thesis was advised by Dr. Ranga Rodigo.

Email  /  Bio  /  Google Scholar  /  Github /  Photos /  Art




Cars don't run like cheetahs, Planes don't fly like birds and Machines won't think in a way same as humans. They will do better. --Richard Feynman

We want AI agents that can discover like we can, not which contain what we have discovered. --Richard Sutton

Machines should be able to understand the world outside our window. That world may change and evolve, but the machines should perceive trees, cars and spaceships not pixels. --Max Wertheimer, Stan Lee

Jathushan2
Research

My research interests lie in the general area of computer vision and deep learning, particularly in learning from videos, deep neural architectures and meta/continual learning.

Blog 1: Research Statement (as of June 2023)

Generative Pretraining from Videos
Jathushan Rajasegaran, Ilija Radosavovic, Yossi Gandelsman, Rahul Ravishankar,
Christoph Feichtenhofer, Jitendra Malik
Coming soon/ project page/ arxiv

We trained LLaMA models up to 1 billion parameters on 1 trillion visual tokens. The resulting model can do diverse tasks from image, video recognition, video tracking, action prediction, and robotics. We also study the scaling properties of these family of models.



Gaussian Masked Autoencoders
Jathushan Rajasegaran, Xinlei Chan, Rulilong Li, Christoph Feichtenhofer,
Shiry Ginosar, Jitendra Malik
Coming soon/ project page/ arxiv

We trained a masked autoencoder with 3D Gaussians as intermindiate representaions. This allows the models to do zero-shot figure-ground segmentation, image layering, edge detection, while performing same as supervised finetuning tasks.



Humanoid Locomotion as Next Token Prediction
Ilija Radosavovic, Bike Zhang, Baifeng Shi, Jathushan Rajasegaran, Sarthak Kamat,
Trevor Darrell, Koushil Sreenath, Jitendra Malik
project page/ arxiv

Real-world humanoid control as a next token prediction problem, akin to predicting the next word in language.



Humans in 4D: Reconstructing and Tracking Humans with Transformers.
Shubham Goel, Georgios Pavlakos, Jathushan Rajasegaran, Angjoo Kanazawa,
Jitendra Malik
International Conference on Computer Vision (ICCV) , 2023  
project page/ arxiv/ code/ demo

A fully "transformerized" deisgn for Human Mesh Recovery achieves improved precision and remarkable robustness for 3D human reconstruction and tracking!.



On the Benefits of 3D Pose and Tracking for Human Action Recognition.
Jathushan Rajasegaran, Georgios Pavlakos, Angjoo Kanazawa,
Christoph Feichtenhofer, Jitendra Malik
Computer Vision and Pattern Recognition (CVPR) , 2023  
project page/ paper/ arxiv/ code/ demo/ poster

Using 3D human reconstruction and tracking to recognize atomic actions in video.



Tracking People by Predicting 3D Appearance, Location and Pose.
Jathushan Rajasegaran, Georgios Pavlakos, Angjoo Kanazawa, Jitendra Malik
Computer Vision and Pattern Recognition (CVPR) , 2022
(Oral Presentation) (Best paper finalist - Top 0.4%)
paper/ arxiv/ project page/ video/ results/ poster/ code

Performing monocular tracking of people by predicting their appearance, pose and location and in 3D.



Tracking People with 3D Representations.
Jathushan Rajasegaran, Georgios Pavlakos, Angjoo Kanazawa, Jitendra Malik
Neural Information Processing Systems (NeurIPS), 2021
paper/ arxiv/ project page/ video/ code/ poster

Performing monocular tracking of people by lifting them to 3D and then using 3D representations of their appearance, pose and location.

iTAML: An Incremental Task-Agnostic Meta-learning Approach.
Jathushan Rajasegaran, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan
Computer Vision and Pattern Recognition (CVPR) , 2020
paper/ arxiv/ slides/ video/ code

By learning generic represenatations from past tasks, we can easily adapt to new tasks as well as remember old tasks.

Random Path Selection for Incremental Learning.
Jathushan Rajasegaran, Munawar Hayat, Salman Khan,
Fahad Shahbaz Khan, Ling Shao
Neural Information Processing Systems (NeurIPS), 2019
paper/ arxiv/ poster/ code

We increase the width of a ResNet like model by adding extra skip connections when new tasks are introduced.

DeepCaps: Going Deeper with Capsule Networks
Jathushan Rajasegaran, Vinoj Jayasundara, Sandaru Jayasekara,
Hirunima Jayasekara, Suranga Seneviratne, Ranga Rodrigo
Computer Vision and Pattern Recognition (CVPR) , 2019   (Oral Presentation)
paper/ poster/ video/ code

Capsule Networks are cool, but they are shallow. We can increase the depth by 3D convolutions and skip connections.

TextCaps: Handwritten Character Recognition with Very Small Datasets
Vinoj Jayasundara, Sandaru Jayasekara, Hirunima Jayasekara, Jathushan Rajasegaran,
Suranga Seneviratne, Ranga Rodrigo
Winter Conference on Applications of Computer Vision (WACV) , 2019
paper/ arxiv/ poster/ code

Capsule Networks can capture actual variations that are present in human hand writing, so we generate more data and retrain the capsule networks.

A Multi-modal Neural Embeddings Approach for Detecting Mobile Counterfeit Apps: A Case Study on Google Play Store
Jathushan Rajasegaran, Naveen Karunanayake, Ashanie Gunathillake,
Suranga Seneviratne, Guillaume Jourjon
International World Wide Web Conference (WWW), 2019
paper/ arxiv/ poster/

We use content and style representations detect counterfeit apps in playstore.


Website source from Jon Barron here