Siddhant Agarwal

I am a final year undergraduate student from the Department of Computer Science and Engineering at the Indian Institute of Technology Kharagpur. I have spent time at the Autonomous Ground Vehicle Research Group at IIT Kharagpur. I have been a part of the Computer Vision and Intelligence Research Lab and worked with Prof. Abir Das on RL and explainable AI for the past two years. I have interned under Prof. Zico Kolter and Prof. Xiang Ren in 2020 and am currently working with Rishabh Agarwal and Prof. Aaron Courville on generalization in reinforcement learning.

Google Scholar / Github / Twitter / LinkedIn/ Email / Resume

profile photo

My primary research interests are reinforcement learning and robot learning. I aim to develop effective data-driven RL algorithms. I want to work at the intersection of offline RL and model-based RL in the areas of generalisability of algorithms and address safety in exploration.

Behavior Predictive Representations for Generalization in Reinforcement Learning
Siddhant Agarwal, Aaron Courville, Rishabh Agarwal
NeurIPS 2021 workshop on Deep Reinforcement Learning and Ecological Theory of RL

We introduce latent representations that can predict the behavior of the agent at future steps to improve generalization in RL.

Reinforcement Explanation Learning
Siddhant Agarwal, Owais Iqbal, Sree Aditya Buridi, Mada Manjusha, Abir Das
NeurIPS 2021 workshop on eXplainable AI approaches for debugging and diagnosis
arXiv / code

We reformulate the process of generating saliency maps using perturbation based methods for black box models as a Markov Decsion Process and use RL to optimally search for the best saliency map, thereby reducing the inference time without hurting the performance.

Poisoned Classifiers are not only backdoored, they are fundamentally broken
Mingjie Sun, Siddhant Agarwal, Zico Kolter
ICLR 2021 workshop on Security and Safety in Machine Learning systems, Under review at ICLR 2022.
arXiv / code

We show that backdoored classifiers can be attacked by anyone rather than only the adversary. We propose an attack that generates alternate triggers for the poisoned classifiers.

Learning to Deceive Knowledge Graph Augmented Models via Targeted Perturbations
Mrigank Raman, Aaron Chan*, Siddhant Agarwal*, Peifeng Wang, Hansen Wang, Sungchul Kim, Ryan Rossi, Handong Zhao, Nedim Lipka, Xiang Ren
* denotes equal contribution
International Conference of Learned Representations 2021 and NeurIPS 2020 workshop on Knowledge Representation and Reasoning in Machine Learning [Best paper nomination]
arXiv / code

We show that using reinforcement learning (or even simple heuristics) we can produce deceptively perturbed knowledge graphs that preserve the downstream performance of the kg-aumented models.

Traffic Sign Classification using HOG-SURF features and Convolutional Neural Networks
Rishabh Madan*, Deepank Agrawal*, Shreyas Kowshik*, Harsh Maheshwari*, Siddhant Agarwal*, Debashish Chakravarty
* denotes equal contribution
International Conference on Pattern Recognition Application and Methods, Prague, 2019

We use a hybrid CNN archticture that uses two image processing features to classify the images. The CNN architecture has significantly less number of parameters than any of the state of the art methods on GTSRB.

Autonomous Ground Vehicles Research Group IIT Kharagpur

AGV is a multidisciplinary research group aimed at building a fully operational self driving car. I worked on various projects involving Motion Planning, Reinforcement Learning and Computer Vision.

Cluster Management System Distribution Systems Project

We studied different design paradigms for developing a distributed cluster management system. Further, we built a simple, fault-tolerant cluster management service using master-slave active-passive topology

Accelerating Graph Algorithms using GPU CUDA Programming Project

We modify the commonly used Graph Algorithms like BFS, DFS, Dijkstra and Flyod Warshall so that they can be executed in a Graphics Processing Unit, bringing about a huge speedup. This required us to parallelize sections of the algorithm into a multi-thread implementation.

Just A Rather Very Interesting Chatbot Software Engineering Project

Build an android chatbot application that identified the emotion of the conversation and generate replies using corresponding seq2seq models trained for specific emotions.

Webpage template courtesy: Jon Barron