This website is best viewed on a desktop browser.
Vaidehi Patil

Hi, I am Vaidehi Patil, a third-year PhD student in the Computer Science Department at the University of North Carolina at Chapel Hill advised by Prof. Mohit Bansal. I am a part of the MURGe-Lab and the broader UNC-NLP group. My research goal is to make deep learning models more safe and responsible for real world applications. For this goal, I work on topics like safety, privacy, security for LLMs, multimodal models, multi-agent systems. I'm also interested in controlling and steering ML models using techniques such as model editing.

I received my Interdisciplinary Dual Degree: B. Tech. in Electrical Engineering and M. Tech. in AI and Data Science and a Minor Degree in Computer Science and Engineering from IIT Bombay in 2022. I have spent wonderful summers as an intern at Adobe Research in Summer 2020 and Summer 2021, at Amazon AGI in Summer 2023 and Apple in Summer 2024. At IIT Bombay, I was advised by Prof. Sunita Sarawagi from IIT Bombay and Partha Talukdar from Google Research India.


Research Interests

  • Natural Language Processing
  • Machine Learning
  • Multimodality
  • AI Safety
  • Privacy
  • Interpretability
  • Multi-agent systems

Education

  • PhD in Computer Science, 2022-present
    University of North Carolina at Chapel Hill
  • M. Tech. in AI and Data Science, 2017-2022
    Indian Institute of Technology Bombay
  • B. Tech. in Electrical Engineering, 2017-2022
    Indian Institute of Technology Bombay
  • Minor in Computer Science and Engineering, 2017-2022
    Indian Institute of Technology Bombay


Recent News




Experience

  • Summer 2024 - Research Intern at Apple
  • Summer 2023 - Research Intern at Amazon AGI
  • Summer 2021 - Research Intern at Adobe Research
  • Summer 2020 - Research Intern at Adobe Research

Publications and Preprints


Publications and Preprints

  • SAFREE: Train-free And Adaptive Guard For Safe Text-to-Image And Video Generation   ICLR 2025
    Jaehong Yoon, Shoubin Yu, Vaidehi Patil, Huaxiu Yao, Mohit Bansal
    Abstract | ArXiv | | Project Webpage

  • UPCORE: Utility-Preserving Coreset Selection for Balanced Unlearning   Preprint
    Vaidehi Patil, Elias Stengel-Eskin, Mohit Bansal
    Abstract | ArXiv |

  • Unlearning Sensitive Information in Multimodal LLMs: Benchmark and Attack-Defense Evaluation   TMLR 2024
    Vaidehi Patil, Yi-Lin Sung, Peter Hase, Jie Peng, Tianlong Chen, Mohit Bansal
    Abstract | ArXiv | | Talk | Slides | Dataset

  • REFINESUMM: Self-Refining MLLM for Generating a Multimodal Summarization Dataset   ACL 2024
    Vaidehi Patil, Leonardo F. R. Ribeiro, Mengwen Liu, Mohit Bansal, Markus Dreyer
    Abstract | ArXiv | Poster | Slides | Talk |

  • Debiasing Multimodal Models via Causal Information Minimization   EMNLP 2023 Findings
    Vaidehi Patil, Adyasha Maharana, Mohit Bansal
    Abstract | ArXiv | Poster | Slides | Talk |

  • Can Sensitive Information Be Deleted From LLMs? Objectives for Defending Against Extraction Attacks   ICLR 2024 Spotlight
    Vaidehi Patil*, Peter Hase*, Mohit Bansal
    Abstract | ArXiv | Poster | Slides | Talk |

  • GEMS: Scene Expansion using Generative Models of Graphs   WACV 2023
    Rishi Agarwal, Tirupati Saketh Chandra, Vaidehi Patil, Aniruddha Mahapatra, Kuldeep Kulkarni, Vishwa Vinay
    Abstract

  • Overlap-based Vocabulary Generation Improves Cross-lingual Transfer Among Related Languages   ACL 2022 Oral
    Vaidehi Patil, Partha Talukdar, Sunita Sarawagi
    Abstract | ArXiv | Poster | Slides | Talk |

  • Detecting Document Versions and Their Ordering In a Collection   WISE 2021 (Best Paper Runner-Up)
    Natwar Modani, Anurag Maurya, Gaurav Verma, Inderjeet Nair, Vaidehi Patil, Anirudh Kanfade
    Abstract | ArXiv |

  • Exploiting Language Relatedness for Low Web-Resource Language Model Adaptation: An Indic Languages Study   ACL 2021
    Yash Khemchandani, Sarvesh Mehtani, Vaidehi Patil, Abhijeet Awasthi, Partha Talukdar, Sunita Sarawagi
    Abstract | ArXiv | Poster | Slides | Talk |





Teaching

Teaching Assistant at IIT Bombay

Misc

Other Professional Activities