Kamalesh Kumar

Hi! I am Kamalesh, a CS grad student at UMass Amherst. I’m advised by Professor Bruno Castro Da Silva who also directs the Autonomous Learning Laboratory. My research interests lie in foundational aspects of Reinforcement Learning (RL), more specifically on continually learning agents, and understanding what existing approaches lack in giving rise to lifelong adaptive behaviour.

Currently, I am also working alongside Professor Chuang Gan at the Embodied-AGI lab, on Pareto-optimal LLM workflow optimization via RL. Previously, I was associated with Resource Bounded Reasoning Lab under Professor Slomo Zilberstein, where I worked on Non-stationarity under the average-reward criterion. In my previous research internships, I worked with Dr Jean-Alexis Delamer and Dr James Hughes on reward function optimization via genetic programming. I have also worked on problems in adversarial and robust RL along with Dr Muni Pydi in the Machine Intelligence and Learning Systems group at Paris Dauphine University. While I was an undergrad at IIT Madras, I worked with the Advanced geometric computing lab under Professor Ramanthan on Wasserstein-GANs.

I am in the job market starting May 2026. More specifically, I am interested about Reinforcement Learning in these two domains: 1. autonomy & robotics 2. post-training & alignment of foundation models. Reach out to me to talk anything about RL (and why it is the key ingredient for AGI 😉).

news

Mar 15, 2026	“RevovleRL” accepted into IEEE World Congress on Computational Intelligence
Sep 3, 2024	Completed my undergrad at IIT Madras and started my masters in CS at UMass Amherst!
Jul 9, 2024	“SketchCleanGAN” accepted into EuroGraphics Symposium on 3D Object Retrieval
Jun 24, 2023	“SketchCADGAN” accepted into EuroGraphics Symposium on 3D Object Retrieval

publications

2024

SketchCleanGAN: A generative network to enhance and correct query sketches for improving 3D CAD model retrieval systems

Kamalesh Kumar Kosalaraman, Prasad Pralhad Kendre, Raghwani Dhaval Manilal, and 1 more author

Published in Elsevier, Computer & Graphics, 2024

Abs

Given an input query, a search and retrieval system fetches relevant information from a dataset. In the Engineering domain, such a system is beneficial for tasks such as design reuse. A two-dimensional (2D) sketch is more conducive for an end user to give as a query than a three-dimensional (3D) object. Such query sketches, nevertheless, will inevitably contain defects like incomplete lines, mesh lines, overdrawn areas, missing areas, etc. Since a retrieval system’s results are only as good as the query, it is necessary to improve the query sketches. In this paper, the problem of transforming a defective CAD sketch into a defect-free sketch is addressed using Generative Adversarial Networks (GANs), which, to the best of our knowledge, has not been investigated before. We first create a dataset of 534 hand-drawn sketches by tracing the boundaries of images of CAD models. We then pair the corrected sketches with their corresponding defective sketches and use them for training a C-WGAN (Conditional Wasserstein Generative Adversarial Network), called SketchCleanGAN. We model the transformation from defective to defect-free sketch as a factorization of the defective input sketch and then translate it to the space of defect-free sketch. We propose a three-branch strategy to this problem. Ablation studies and comparisons with other state-of-the-art techniques demonstrate the efficacy of the proposed technique. Additionally, we also contribute to a dataset of around 58000 improved sketches using the proposed framework.

2023

SketchCADGAN: A generative approach for completing partially drawn query sketches of engineering shapes to enhance retrieval system performance

Prasad Pralhad Kendre, Kamalesh Kumar Kosalaraman, Sanjay Santhosh Kumar Jayasree, and 3 more authors

Published in Elsevier, Computer & Graphics, 2023

Abs Bib paper

Retrieval systems are commonly used to find relevant data in large datasets. In engineering, these systems are useful for locating specific engineering shapes in a large dataset of engineering components. When end users want to search for a shape, they prefer a two-dimensional (2D) sketch over a three-dimensional (3D) object. However, users lacking domain knowledge may struggle to generate a complete query sketch and provide a partially completed sketch instead. Retrieving relevant information from partially drawn sketches is difficult because they may have missing edges, partially drawn circles, holes, ovals, etc. Most retrieval systems compare the similarity between the query and items in the database, so incomplete sketches may be ineffective in finding the relevant information. To address this problem with incomplete sketches, we propose a new generative adversarial network called SketchCADGAN. This network uses a two-stage cascaded architecture, with the first network attempting to predict a CAD model image from an incomplete sketch and the second network using the CAD model image to predict a completed sketch. Both networks are trained together adversarially. Our approach is proven more effective than other advanced techniques through qualitative and quantitative comparisons. Furthermore, we present the results of the retrieval system using both partially drawn and completed sketches, and demonstrate that incorporating completed sketches from the suggested cascaded GAN architecture results in improved retrieval performance.
@article{KENDRE202355, title = {SketchCADGAN: A generative approach for completing partially drawn query sketches of engineering shapes to enhance retrieval system performance}, journal = {Computers & Graphics}, volume = {115}, pages = {55-68}, year = {2023}, issn = {0097-8493}, doi = {https://doi.org/10.1016/j.cag.2023.06.028}, author = {Kendre, Prasad Pralhad and Kosalaraman, Kamalesh Kumar and Jayasree, Sanjay Santhosh Kumar and Rajan, Sreehari and Jayan, Akash and Muthuganapathy, Ramanathan}, keywords = {Search and retrieval, Generative adversarial networks(GANs), Sketch completion, Sketch inpainting, CAD model sketches, Incomplete sketches}, }