Aniruddha Saha

(awe-knee-rue-though sha-haa)

Experienced in AI/ML Security, Safety, and Trustworthiness

ani0075saha [at] gmail [dot] com

I am on the job market.
Please reach out, if you think a research opportunity (academia or industry) would be a good fit for me.

News
Bio

I am currently working as the Lead AI Engineer at an early-stage startup, building a product for video content understanding.

Most recently, I was a Postdoctoral Associate at the Center for Machine Learning (CML) in University of Maryland, College Park. I was advised by Tom Goldstein.

I received my PhD in Computer Science at University of Maryland, Baltimore County. I was advised by Hamed Pirsiavash. In my dissertation, I studied ways in which state-of-the-art deep learning methods for computer vision are vulnerable to backdoor attacks and proposed defense methods to remedy the vulnerabilities.

During my PhD, I have worked as a Machine Learning Research Intern at Bosch Center for AI, an Applied Scientist Intern at Amazon Rekognition, and a Machine Learning Intern at Matroid.

Prior to this, I was a Software Engineer at Samsung Research Institute Bangalore, India where I was part of the DRAM Group of Samsung Semiconductor India Research.

My hobbies include photography, writing, playing football and chess. I support Manchester United FC. See what I am currently reading.

Publications

Complete list on Google Scholar.

Generating Potent Poisons and Backdoors from Scratch with Guided Diffusion Paper

Hossein Souri, Arpit Bansal, Hamid Kazemi, Liam Fowl, Aniruddha Saha, Jonas Geiping, Andrew Gordon Wilson, Rama Chellappa, Tom Goldstein, Micah Goldblum

ICML 2024 Workshop on the Next Generation of AI Safety (Oral)

A Closer Look at Robustness of Vision Transformers to Backdoor Attacks Paper

Akshayvarun Subramanya, Aniruddha Saha*, Soroush Abbasi Koohpayegani*, Ajinkya Tejankar, Hamed Pirsiavash

WACV 2024

*equal contribution

Revisiting Image Classifier Training for Improved Certified Robust Defense against Adversarial Patches Paper

Aniruddha Saha*, Shuhua Yu*, Mohammad Sadegh Norouzzadeh, Wan-Yi Lin, Chaithanya Kumar Mummadi

Transactions on Machine Learning Research (TMLR) Oct 2023

*equal contribution

Backdoor Attacks on Self-Supervised Learning Paper Slides Code

Aniruddha Saha, Ajinkya Tejankar, Soroush Abbasi Koohpayegani, Hamed Pirsiavash

CVPR 2022 (Oral)

Role of Spatial Context in Adversarial Robustness for Object Detection Paper Slides Video Code

Aniruddha Saha*, Akshayvarun Subramanya*, Koninika Patil, Hamed Pirsiavash

CVPR 2020 Workshop on Adversarial Machine Learning in Computer Vision (Long Paper)

*equal contribution

Universal Litmus Patterns: Revealing Backdoor Attacks in CNNs Paper Webpage Slides Video Code

Soheil Kolouri*, Aniruddha Saha*, Hamed Pirsiavash, Heiko Hoffmann

CVPR 2020 (Oral)

*equal contribution

Hidden Trigger Backdoor Attacks Paper Slides Poster Code

Aniruddha Saha, Akshayvarun Subramanya, Hamed Pirsiavash

AAAI 2020 (Oral)

Bring Your Own Data! Self-Supervised Evaluation for Large Language Models Paper

Neel Jain*, Khalid Saifullah*, Yuxin Wen, John Kirchenbauer, Manli Shu, Aniruddha Saha, Micah Goldblum, Jonas Geiping, Tom Goldstein

First Conference on Language Modeling (COLM) 2024

*equal contribution

Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text Paper

Abhimanyu Hans*, Avi Schwarzschild*, Valeriia Cherepanova, Hamid Kazemi, Aniruddha Saha, Micah Goldblum, Jonas Geiping, Tom Goldstein

ICML 2024

*equal contribution

NEFTune - Noisy Embeddings Improve Instruction Finetuning Paper

Neel Jain*, Ping-yeh Chiang*, Yuxin Wen*, John Kirchenbauer, Hong-Min Chu, Gowthami Somepalli, Brian R Bartoldson, Bhavya Kailkhura, Avi Schwarzschild, Aniruddha Saha, Micah Goldblum, Jonas Geiping, Tom Goldstein

ICLR 2024

*equal contribution

On the Reliability of Watermarks for Large Language Models Paper

John Kirchenbauer*, Jonas Geiping*, Yuxin Wen, Manli Shu, Khalid Saifullah, Kezhi Kong, Kasun Fernando, Aniruddha Saha, Micah Goldblum, Tom Goldstein

ICLR 2024

*equal contribution

An Adaptive Foreground-Background Separation Method for Effective Binarization of Document Images Paper

Bishwadeep Das, Showmik Bhowmik, Aniruddha Saha, Ram Sarkar

Proceedings of the Eighth International Conference on Soft Computing and Pattern Recognition (SoCPaR 2016)

Baseline Defenses for Adversarial Attacks Against Aligned Language Models Paper

Neel Jain, Avi Schwarzschild, Yuxin Wen, Gowthami Somepalli, John Kirchenbauer, Ping-yeh Chiang, Micah Goldblum, Aniruddha Saha, Jonas Geiping, Tom Goldstein

*equal contribution

Timeline
Talks
Service & Leadership
Program Committee (Reviewer)

Conferences:
ICLR 2025, AAAI 2025
NeurIPS 2024 Ethics, ICLR 2024
NeurIPS 2023, ICCV 2023, CVPR 2023
ECCV 2022, ICPR 2022, CVPR 2022
ICCV 2021*
ICPR 2020

Workshops:
CVPR 2022, ICLR 2022
ICCV 2021, ICML 2021, CVPR 2021, ICLR 2021
ECCV 2020, CVPR 2020, ICLR 2020

Journals: IEEE TPAMI, IEEE TIFS, IEEE TETCI, IET Computer Vision

*Outstanding Reviewer Award

Program Chair (Organizer)
Backdoors in Deep Learning Workshop at NeurIPS 2023.
Boundless Fun

Birding (June 21, 2024)
I am fascinated by the natural world. I have been lucky to have unbounded leisure time during my recent job search. I used this opportunity to put my photography and bird watching interests together. Here are some results. Please reach out with kind and encouraging words if you love them. I would be happy to give you tips if you have also kept this interest under wraps for a long time. It doesn't take much, but provides limitless rewards.

UEFA Euros 2024 (June 14, 2024)
I am ready for a month of top quality football. I have had a soft-spot for Portugal since I saw the team, and Ronaldo at the 2004 Euros. Greece caused a major upset.

Acknowledgement

This website uses the website design and template by Martin Saveski.

Aniruddha Saha
Last updated 10/10/2024