Home
About me
Publications

Detailed Publications

Continual Learning

  • Dipam Goswami, Albin Soutif–Cormerais, Yuyang Liu, Sandesh Kamath, Bartłomiej Twardowski, Joost van de Weijer, Resurrecting Old Classes with New Data for Exemplar-Free Continual Learning, IEEE/CVF Conference on Computer Vision and Pattern Recognition, (CVPR’24). html pdf

  • Sandesh Kamath, Albin Soutif-Cormerais, Joost van de Weijer, Bogdan Raducanu, The Expanding Scope of the Stability Gap: Unveiling its Presence in Joint Incremental Learning of Homogeneous Tasks, CVPR 2024 Workshop on ‘‘Continual Learning in Computer Vision’’, (CVPR-W’24). html pdf

Generative AI

  • Xide Xu, Sandesh Kamath, Muhammad Atif Butt, Bogdan Raducanu, An h-space Based Adversarial Attack for Protection Against Few-shot Personalization, ACM Multimedia, (ACM MM’25). pdf

  • Xide Xu, Muhammad Atif Butt, Sandesh Kamath, Bogdan Raducanu, Privacy Protection in Personalized Diffusion Models via Targeted Cross-Attention Adversarial Attack - NeurIPS 2024 Workshop on ‘‘Safe Generative AI’’.

Explainable AI

  • Sandesh Kamath, Sankalp Mittal, Amit Deshpande, Vineeth N Balasubramanian, Rethinking Robustness of Model Attributions, The 38th Annual AAAI Conference on Artificial Intelligence, (AAAI’24). html pdf code

  • Sandesh Kamath, Amlan Jyoti, Karthik Balaji Ganesh, Manoj Gayala, Nandita Lakshmi Tunuguntla, Vineeth N Balasubramanian, On the Robustness of Explanations of Deep Neural Network Models: A Survey pdf

Adversarial Robustness

  • Sandesh Kamath, Amit Deshpande, K V Subrahmanyam, Vineeth N Balasubramanian, Can we have it all? On the Trade-off between Spatial and Adversarial Robustness of Neural Networks, Conference on Neural Information Processing Systems (NeurIPS’21) pg 27462-27474. html pdf code

  • Sandesh Kamath, Amit Deshpande, K V Subrahmanyam, Vineeth N Balasubramanian, Universalization of any adversarial attack using very few test examples - CODS-COMAD 2022 (Best Paper Award, Research Track), pages 72-80. html pdf code

  • Sandesh Kamath, Amit Deshpande, K V Subrahmanyam, On Adversarial Robustness of Small vs. Large Batch Training - ICML 2019 Workshop on ‘‘Understanding and Improving Generalization in Deep Learning’’. html pdf

  • Sandesh Kamath, Amit Deshpande, K V Subrahmanyam, Better Generalization with Adaptive Adversarial Training - ICML 2019 Workshop on ‘‘Understanding and Improving Generalization in Deep Learning’’. html pdf

  • Sandesh Kamath, Amit Deshpande, Understanding Adversarial Robustness of Symmetric Networks - ICML 2018 Workshop on ‘‘Towards learning with limited labels: Equivariance, Invariance, and Beyond’’. html

  • Sandesh Kamath, Amit Deshpande, K V Subrahmanyam, On Universalized Adversarial and Invariant Perturbation, arxiv.2006.04449, 2020. pdf.

  • Sandesh Kamath, Amit Deshpande, K V Subrahmanyam, How do SGD hyperparameters in natural training affect adversarial robustness? arxiv.2006.11604, 2020. pdf.