Amartya Sanyal

73 papers A* 19A 3Journal 47Unranked 3

Year	Rank	Type	Title / Venue / Authors
2026	J	`jnl`	Adaptive Sampling for Private Worst-Case Group Optimization. CoRR Max Cairney-Leeming, Amartya Sanyal, Christoph H. Lampert
2026	J	`jnl`	Delta-Influence: Identifying Poisons via Influence Functions. Trans. Mach. Learn. Res. Wenjie Li, Jiawei Li, Pengcheng Zeng, Christian Schroeder de Witt, Ameya Prabhu, Amartya Sanyal
2026	J	`jnl`	Language Generation with Replay: A Learning-Theoretic View of Model Collapse. CoRR Giorgio Racca, Michal Valko, Amartya Sanyal
2026	A*	`conf`	Learning in an Echo Chamber: Online Learning with Replay Adversary. SODA Daniil Dmitriev, Harald Eskelund Franck, Carolin Heinzler, Amartya Sanyal
2026	J	`jnl`	Less Noise, Same Certificate: Retain Sensitivity for Unlearning. CoRR Carolin Heinzler, Kasra Malihi, Amartya Sanyal
2026	J	`jnl`	LoRA and Privacy: When Random Projections Help (and When They Don't). CoRR Yaxi Hu, Johanna Düngler, Bernhard Schölkopf, Amartya Sanyal
2025	A	`conf`	Accuracy on the wrong line: On the pitfalls of noisy data for out-of-distribution generalisation. AISTATS Amartya Sanyal, Yaxi Hu, Yaodong Yu, Yian Ma, Yixin Wang, Bernhard Schölkopf
2025	J	`jnl`	An Iterative Algorithm for Differentially Private CoRR Johanna Düngler, Amartya Sanyal
2025	A*	`conf`	Differentially Private Steering for Large Language Model Alignment. ICLR Anmol Goel, Yaxi Hu, Iryna Gurevych, Amartya Sanyal
2025	J	`jnl`	Differentially Private Steering for Large Language Model Alignment. CoRR Anmol Goel, Yaxi Hu, Iryna Gurevych, Amartya Sanyal
2025	J	`jnl`	Fairness for the People, by the People: Minority Collective Action. CoRR Omri Ben-Dov, Samira Samadi, Amartya Sanyal, Alexandru Tifrea
2025	J	`jnl`	Learning in an Echo Chamber: Online Learning with Replay Adversary. CoRR Daniil Dmitriev, Harald Eskelund Franck, Carolin Heinzler, Amartya Sanyal
2025	J	`jnl`	Online Learning and Unlearning. CoRR Yaxi Hu, Bernhard Schölkopf, Amartya Sanyal
2025	J	`jnl`	Open Problems in Machine Unlearning for AI Safety. CoRR Fazl Barez, Tingchen Fu, Ameya Prabhu, Stephen Casper, Amartya Sanyal, Adel Bibi, Aidan O'Gara, Robert Kirk, Ben Bucknall, Timothy Fist, Luke Ong, Philip Torr, Kwok-Yan Lam, Robert Trager, David Krueger, Sören Mindermann, José Hernández-Orallo, Mor Geva, Yarin Gal
2025	A*	`conf`	Protecting against simultaneous data poisoning attacks. ICLR Neel Alex, Shoaib Ahmed Siddiqui, Amartya Sanyal, David Krueger
2025	A*	`conf`	Provable unlearning in topic modeling and downstream tasks. ICLR Stanley Wei, Sadhika Malladi, Sanjeev Arora, Amartya Sanyal
2024	J	`jnl`	Accuracy on the wrong line: On the pitfalls of noisy data for out-of-distribution generalisation. CoRR Amartya Sanyal, Yaxi Hu, Yaodong Yu, Yian Ma, Yixin Wang, Bernhard Schölkopf
2024	A	`conf`	Certified private data release for sparse Lipschitz functions. AISTATS Konstantin Donhauser, Johan Lokna, Amartya Sanyal, March Boedihardjo, Robert Hönig, Fanny Yang
2024	J	`jnl`	Corrective Machine Unlearning. CoRR Shashwat Goel, Ameya Prabhu, Philip Torr, Ponnurangam Kumaraguru, Amartya Sanyal
2024	J	`jnl`	Corrective Machine Unlearning. Trans. Mach. Learn. Res. Shashwat Goel, Ameya Prabhu, Philip Torr, Ponnurangam Kumaraguru, Amartya Sanyal
2024	J	`jnl`	Delta-Influence: Unlearning Poisons via Influence Functions. CoRR Wenjie Li, Jiawei Li, Christian Schröder de Witt, Ameya Prabhu, Amartya Sanyal
2024	A*	`conf`	On the Growth of Mistakes in Differentially Private Online Learning: A Lower Bound Perspective. COLT Daniil Dmitriev, Kristóf Szabó, Amartya Sanyal
2024	J	`jnl`	On the Growth of Mistakes in Differentially Private Online Learning: A Lower Bound Perspective. CoRR Daniil Dmitriev, Kristóf Szabó, Amartya Sanyal
2024	—	`conf`	PILLAR: How to make semi-private learning more effective. SaTML Francesco Pinto, Yaxi Hu, Fanny Yang, Amartya Sanyal
2024	J	`jnl`	Protecting against simultaneous data poisoning attacks. CoRR Neel Alex, Shoaib Ahmed Siddiqui, Amartya Sanyal, David Krueger
2024	A*	`conf`	Provable Privacy with Non-Private Pre-Processing. ICML Yaxi Hu, Amartya Sanyal, Bernhard Schölkopf
2024	J	`jnl`	Provable Privacy with Non-Private Pre-Processing. CoRR Yaxi Hu, Amartya Sanyal, Bernhard Schölkopf
2024	J	`jnl`	Provable unlearning in topic modeling and downstream tasks. CoRR Stanley Wei, Sadhika Malladi, Sanjeev Arora, Amartya Sanyal
2024	A*	`conf`	Robust Mixture Learning when Outliers Overwhelm Small Groups. NeurIPS Daniil Dmitriev, Rares-Darius Buhai, Stefan Tiegel, Alexander Wolters, Gleb Novikov, Amartya Sanyal, David Steurer, Fanny Yang
2024	J	`jnl`	Robust Mixture Learning when Outliers Overwhelm Small Groups. CoRR Daniil Dmitriev, Rares-Darius Buhai, Stefan Tiegel, Alexander Wolters, Gleb Novikov, Amartya Sanyal, David Steurer, Fanny Yang
2024	A*	`conf`	The Role of Learning Algorithms in Collective Action. ICML Omri Ben-Dov, Jake Fawkes, Samira Samadi, Amartya Sanyal
2024	J	`jnl`	The Role of Learning Algorithms in Collective Action. CoRR Omri Ben-Dov, Jake Fawkes, Samira Samadi, Amartya Sanyal
2024	A*	`conf`	What Makes and Breaks Safety Fine-tuning? A Mechanistic Study. NeurIPS Samyak Jain, Ekdeep Singh Lubana, Kemal Oksuz, Tom Joy, Philip Torr, Amartya Sanyal, Puneet K. Dokania
2024	J	`jnl`	What Makes and Breaks Safety Fine-tuning? A Mechanistic Study. CoRR Samyak Jain, Ekdeep Singh Lubana, Kemal Oksuz, Tom Joy, Philip H. S. Torr, Amartya Sanyal, Puneet K. Dokania
2023	A*	`conf`	A law of adversarial risk, interpolation, and label noise. ICLR Daniel Paleka, Amartya Sanyal
2023	A*	`conf`	Can semi-supervised learning use all the data effectively? A lower bound perspective. NeurIPS Alexandru Tifrea, Gizem Yüce, Amartya Sanyal, Fanny Yang
2023	J	`jnl`	Can semi-supervised learning use all the data effectively? A lower bound perspective. CoRR Alexandru Tifrea, Gizem Yüce, Amartya Sanyal, Fanny Yang
2023	J	`jnl`	Catastrophic overfitting can be induced with discriminative non-robust features. Trans. Mach. Learn. Res. Guillermo Ortiz-Jiménez, Pau de Jorge, Amartya Sanyal, Adel Bibi, Puneet K. Dokania, Pascal Frossard, Grégory Rogez, Philip Torr
2023	A*	`conf`	Certifying Ensembles: A General Certification Theory with S-Lipschitzness. ICML Aleksandar Petrov, Francisco Eiras, Amartya Sanyal, Philip H. S. Torr, Adel Bibi
2023	J	`jnl`	Certifying Ensembles: A General Certification Theory with S-Lipschitzness. CoRR Aleksandar Petrov, Francisco Eiras, Amartya Sanyal, Philip H. S. Torr, Adel Bibi
2023	J	`jnl`	How robust accuracy suffers from certified training with convex relaxations. CoRR Piersilvio De Bartolomeis, Jacob Clarysse, Amartya Sanyal, Fanny Yang
2023	A*	`conf`	How robust is unsupervised representation learning to distribution shift? ICLR Yuge Shi, Imant Daunhawer, Julia E. Vogt, Philip H. S. Torr, Amartya Sanyal
2023	J	`jnl`	PILLAR: How to make semi-private learning more effective. CoRR Francesco Pinto, Yaxi Hu, Fanny Yang, Amartya Sanyal
2023	J	`jnl`	Sample-efficient private data release for Lipschitz functions under sparsity assumptions. CoRR Konstantin Donhauser, Johan Lokna, Amartya Sanyal, March Boedihardjo, Robert Hönig, Fanny Yang
2022	J	`jnl`	A law of adversarial risk, interpolation, and label noise. CoRR Daniel Paleka, Amartya Sanyal
2022	J	`jnl`	Catastrophic overfitting is a bug but also a feature. CoRR Guillermo Ortiz-Jiménez, Pau de Jorge, Amartya Sanyal, Adel Bibi, Puneet K. Dokania, Pascal Frossard, Grégory Rogez, Philip H. S. Torr
2022	J	`jnl`	Do you pay for Privacy in Online learning? CoRR Amartya Sanyal, Giorgia Ramponi
2022	J	`jnl`	How robust are pre-trained models to distribution shift? CoRR Yuge Shi, Imant Daunhawer, Julia E. Vogt, Philip H. S. Torr, Amartya Sanyal
2022	J	`jnl`	How unfair is private learning ? CoRR Amartya Sanyal, Yaxi Hu, Fanny Yang
2022	A	`conf`	How unfair is private learning? UAI Amartya Sanyal, Yaxi Hu, Fanny Yang
2022	A*	`conf`	Make Some Noise: Reliable and Efficient Single-Step Adversarial Training. NeurIPS Pau de Jorge Aranda, Adel Bibi, Riccardo Volpi, Amartya Sanyal, Philip H. S. Torr, Grégory Rogez, Puneet K. Dokania
2022	J	`jnl`	Make Some Noise: Reliable and Efficient Single-Step Adversarial Training. CoRR Pau de Jorge, Adel Bibi, Riccardo Volpi, Amartya Sanyal, Philip H. S. Torr, Grégory Rogez, Puneet K. Dokania
2021	A*	`conf`	How Benign is Benign Overfitting ? ICLR Amartya Sanyal, Puneet K. Dokania, Varun Kanade, Philip H. S. Torr
2021	J	`jnl`	Identifying and Exploiting Structures for Reliable Deep Learning. CoRR Amartya Sanyal
2021	—	`—`	Identifying and exploiting structures for reliable deep learning Amartya Sanyal
2021	A*	`conf`	Progressive Skeletonization: Trimming more fat from a network at initialization. ICLR Pau de Jorge, Amartya Sanyal, Harkirat S. Behl, Philip H. S. Torr, Grégory Rogez, Puneet K. Dokania
2020	A*	`conf`	Calibrating Deep Neural Networks using Focal Loss. NeurIPS Jishnu Mukhoti, Viveka Kulharia, Amartya Sanyal, Stuart Golodetz, Philip H. S. Torr, Puneet K. Dokania
2020	J	`jnl`	Calibrating Deep Neural Networks using Focal Loss. CoRR Jishnu Mukhoti, Viveka Kulharia, Amartya Sanyal, Stuart Golodetz, Philip H. S. Torr, Puneet K. Dokania
2020	J	`jnl`	Hierarchical discovery of large-scale and focal copy number alterations in low-coverage cancer genomes. BMC Bioinform. Ahmed Ibrahim S. Khalil, Costerwell Khyriem, Anupam Chattopadhyay, Amartya Sanyal
2020	J	`jnl`	How benign is benign overfitting? CoRR Amartya Sanyal, Puneet K. Dokania, Varun Kanade, Philip H. S. Torr
2020	J	`jnl`	Identification and utilization of copy number information for correcting Hi-C contact map of cancer cell lines. BMC Bioinform. Ahmed Ibrahim S. Khalil, Siti Rawaidah Binte Mohammad Muzaki, Anupam Chattopadhyay, Amartya Sanyal
2020	J	`jnl`	Progressive Skeletonization: Trimming more fat from a network at initialization. CoRR Pau de Jorge, Amartya Sanyal, Harkirat S. Behl, Philip H. S. Torr, Grégory Rogez, Puneet K. Dokania
2020	A*	`conf`	Stable Rank Normalization for Improved Generalization in Neural Networks and GANs. ICLR Amartya Sanyal, Philip H. S. Torr, Puneet K. Dokania
2019	J	`jnl`	Stable Rank Normalization for Improved Generalization in Neural Networks and GANs. CoRR Amartya Sanyal, Philip H. S. Torr, Puneet K. Dokania
2018	J	`jnl`	Low Rank Structure of Learned Representations. CoRR Amartya Sanyal, Varun Kanade, Philip H. S. Torr
2018	J	`jnl`	Optimizing Non-decomposable Measures with Deep Networks. CoRR Amartya Sanyal, Pawan Kumar, Purushottam Kar, Sanjay Chawla, Fabrizio Sebastiani
2018	J	`jnl`	Optimizing non-decomposable measures with deep networks. Mach. Learn. Amartya Sanyal, Pawan Kumar, Purushottam Kar, Sanjay Chawla, Fabrizio Sebastiani
2018	A*	`conf`	TAPAS: Tricks to Accelerate (encrypted) Prediction As a Service. ICML Amartya Sanyal, Matt J. Kusner, Adrià Gascón, Varun Kanade
2018	J	`jnl`	TAPAS: Tricks to Accelerate (encrypted) Prediction As a Service. CoRR Amartya Sanyal, Matt J. Kusner, Adrià Gascón, Varun Kanade
2017	J	`jnl`	Agent based simulation of the evolution of society as an alternate maximization problem. CoRR Amartya Sanyal, Sanjana Garg, Asim Unmesh
2017	—	`conf`	Agent based simulation of the evolution of society as an alternate maximzation problem. BESC Amartya Sanyal, Sanjana Garg, Asim Unmesh, Harish Karnick
2017	J	`jnl`	Multiscale sequence modeling with a learned dictionary. CoRR Bart van Merriënboer, Amartya Sanyal, Hugo Larochelle, Yoshua Bengio
2016	—	`conf`	A Hybrid Deep Architecture for Face Recognition in Real-Life Scenario. ICVGIP Workshops Amartya Sanyal, Ujjwal Bhattacharya, Swapan K. Parui