Vinu's Homepage

About Me

[CV] [Google Scholar] [Twitter] [LinkedIn]

I am on the job market and actively looking for research positions. Please reach out to me about potential opportunities in your team.

I am a final year Computer Science PhD candidate at The University of Maryland, College Park advised by Prof. Soheil Feizi. I am a recipient of the Kulkarni Fellowship 2023. My research interest broadly lies in the area of Security and Privacy in AI. Currently, I am working as a full-time Student Researcher at Google DeepMind on jailbreaking multimodal AI models. Previously, I worked as a Research Scientist intern at Meta FAIR in Paris on AI watermarking. My PhD thesis at UMD majorly focuses on AI robustness, detectability, and user privacy.

I received the prestigious Director’s Silver Medal and my Bachelor’s degree in Computer Science & Engineering in 2020 from IIT Gandhinagar. I was a Junior Research Fellow with Prof. Anirban Dasgupta at IIT Gandhinagar, India in the Data Science Lab. Before my PhD, I had interned at Caltech, Microsoft Research India, and IISc.

News

[Nov 2024] Invited for a talk at UK AI Safety Institute.
[Nov 2024] Invited for MLOps Podcast.
[Sep 2024] LLM-Check is accepted to NeurIPS 2024.
[Sep 2024] Joining Google DeepMind, Mountain View as a Full-time Student Researcher.
[May 2024] Joining Meta FAIR, Paris as a Research Scientist Intern.
[May 2024] BEAST is accepted to ICML 2024.
[May 2024] Invited talk on AI content detection at US Securities & Exchange Commission!
[Apr 2024] Invited talk on BEAST (jailbreaking) at Amazon AWS Responsible AI team.
[Feb 2024] Interviewed by IEEE Spectrum for expert opinion on AI-text detection.
[Nov 2023] Gave a research talk at Google Research on our work on AI-text detection.

Show more

[Oct 2023]</b> Our work on limitations of AI-image detectors is featured in Wired!
[Jun 2023] I am awarded the Kulkarni Summer Research Fellowship for 2023!
[Jun 2023] Interviewed by New Scientist for expert opinion on model collapse.
[Jun 2023] Our work on hardness of AI-detectors is featured in Washington Post!
[Apr 2023] Recognized as notable reviewer at ICLR 2023.
[Mar 2023] Our work on hardness of AI-detectors is featured in New Scientist!
[Mar 2023] Our work on hardness of AI-detectors is featured in The Register!
[Mar 2023] CUDA is on arXiv.
[Feb 2023] CUDA accepted at CVPR 2023.
[Nov 2022] Reviewer for ICLR 2023.
[Jul 2022] Reviewer for NeurIPS 2022.
[Aug 2021] Excited to join UMD CS for my PhD.
[Jun 2021] Curriculum learning work accepted (spotlight) at SubSetML, ICML 2021.
[Feb 2021] Curriculum learning work preprint now available on arXiv.
[Feb 2021] Reviewer for ICML 2021.
[Sep 2020] Received cash award for CS publication from IIT Gandhinagar.
[Aug 2020] Started working as JRF at IIT Gandhinagar.
[Aug 2020] Received the Director's Silver Medal and B.Tech in CSE.
[Sep 2019] Special mention for poster at UGRC 2019 at IIT Gandhinagar for work at Caltech.
[Sep 2019] Work on Shallow RNN accepted at NeurIPS 2019.
[May 2019] Started working as undergraduate research fellow at Caltech.
[Feb 2019] Work on endoscopy abnormality classification accepted at IEEE ISBI 2019.
[Jan 2019] Started working as research intern at MSR India.
[Nov 2018] Received Caltech's SURF for summer 2019.
[May 2017] Started working as research intern at IISc Bangalore.

My Research

2024

LLM-Check: Investigating Detection of Hallucinations in Large Language Models
G Sriramanan, S Bharti, VS Sadasivan, S Saha, P Kattakinda, S Feizi
[PDF] NeurIPS 2024.

DREW: Towards Robust Data Provenance by Leveraging Error-Controlled Watermarking
M Saberi, VS Sadasivan, A Zarei, H Mahdavifar, S Feizi
[PDF] Preprint Jun 2024.

Fast Adversarial Attacks on Language Models In One GPU Minute
VS Sadasivan, S Saha*, G Sriramanan*, P Kattakinda, A Chegini, S Feizi
[PDF] ICML 2024.
Media 📢: The Register
Talks 🎤: Amazon AWS Responsible AI

Robustness of AI-Image Detectors: Fundamental Limits and Practical Attacks
M Saberi, VS Sadasivan, K Rezaei, A Kumar, A Chegini, W Wang, S Feizi
[PDF] ICLR 2024.
Media 📢: Wired, Bloomberg, The Verge, MIT Technology Review, The Register, UMD CS

2023

Can AI-Generated Text be Reliably Detected?
VS Sadasivan, A Kumar, S Balasubramanian, W Wang, S Feizi
[PDF] Preprint Mar 2023.
Media 📢: Washington Post, Wired, New Scientist, The Register [1], [2], TechSpot, UMD Science
Talks 🎤: US Securities and Exchange Commission, Google Research

Exploring Geometry of Blind Spots in Vision Models
S Balasubramanian*, G Sriramanan*, VS Sadasivan, S Feizi
[PDF] NeurIPS 2023 (Spotlight).

Provable Robustness for Streaming Models with a Sliding Window.
A Kumar, VS Sadasivan, S Feizi
[PDF] Preprint Mar 2023.

CUDA: Convolution-based Unlearnable Datasets.
VS Sadasivan, M Soltanolkotabi, S Feizi
[PDF] IEEE/CVF CVPR 2023.

2021

Statistical measures for defining curriculum scoring function.
VS Sadasivan, A Dasgupta
[PDF] SubSetML Workshop @ ICML 2021 (Spotlight).

2019

Shallow RNN: accurate time-series classification on resource constrained devices.
D Dennis, DAE Acar, V Mandikal, VS Sadasivan, V Saligrama, HV Simhadri, P Jain
[PDF] NeurIPS 2019.

High accuracy patch-level classification of wireless capsule endoscopy images using a convolutional neural network.
VS Sadasivan, CS Seelamantula
[PDF] IEEE ISBI 2019.

Personal Stuff

I come from the beautiful town Kollam in Kerala. In my free time, I love to sing and play the ukulele and guitar. Check out my YouTube and Instagram pages for my musical covers. I also love playing badminton, soccer, ultimate frisbee, and cricket.

Template from this.