Dhruvesh Patel

dp.jpg

dhruveshpate@umass.edu

🌟 I’m looking for internship opportunities for Spring and Summer 2025. Here is my CV. 🌟

I am currently a fourth-year Computer Science PhD student at UMass Amherst working with Prof. Andrew McCallum alongside some amazing colleagues at the Information Extraction and Synthesis Laboratory. I completed my undergraduate at IIT Madras, where I worked on Robotics research, mentored by Prof. Bandyopadhyay.

Outside of my academic pursuits, I’ve been fortunate to have worked with some amazing collaborators from the industry. I have worked as a research scientist inter at Meta Reality Labs and Abridge AI. Before beginning my master’s program at UMass, I worked for two years as a software engineer at MathWorks. I also dedicated a year to collaborating with Prof. Partha Talukdar on solving various NLP problems in the industry.

research

Autoregressive models dominate the scene for generative modeling of non-ordinal discrete data, like text, mostly due to the scalability of pre-training. However, as generative models they have many limitations: limited conditioning and control at inference time, inefficient use of inference time computation by tying the sequence length to the computation, and inability to support non-sequential forms of interaction like edits or deletions. I’m interested in scaling non-autoregessive models like discrete diffusion and flows for text generation either by adapting pre-trained AR models through continued training or by making non-AR pre-training more efficient.

Prior to this, I have worked on non-Euclidean representation learning, energy-based models for discrete data, and compositional generalization in-context learning.

affiliations and internships

news

Oct 1, 2024 Learning Representations for Hierarchies with Minimal Support was accepted at NeurIPS 2024!
Apr 1, 2024 Language Guided Exploration for RL Agents in Text Environments was accepted at NAACL (findings) 2024.
Aug 1, 2023 My work on Pre-trained language models for Visual Planning for Human Assistance, done as a research intern at Meta Reality Labs., has been accepted at ICCV 2023.
Sep 7, 2022 Super excited to start my internship at Meta Reality Labs!
Apr 25, 2022 Excited to present our work on multi-label classification using box embeddings at ICLR 2022!

mentors and collaborators

I have been fortunate to work have worked with many amazing people over the years. Here is a list of my current and previous collaborators.
Michael Boratko [Google] (2019 - 2025)
Tahira Naseem [IBM] (2023 - 2023)
Akash Srivastava [MIT-IBM Research] (2023 - 2023)
Keerthiram Murugesan [IBM] (2022 - 2023)
Kenneth Clarkson [IBM] (2023 - 2023)
Kartik Talamadupula [IBM] (2019 - 2019)
Pavan Kapanipathi [IBM] (2019 - 2019)
Jay-Yoon Lee [Seoul National University] (2020 - 2022)
Partha Talukdar [IISc Bangalore/Google Research] (2018 - 2018)
Sandipan Bandyopadhyay [IIT Madras] (2016 - 2016)
Ruta Desai [Meta AI] (2022 - 2023)
Unnat Jain [Meta AI] (2023 - 2023)

selected publications

2024

  1. Language Guided Exploration for RL Agents in Text Environments
    Hitesh Golchha, Sahil Yerawar, Dhruvesh Patel, Soham Dan, and Keerthiram Murugesan
    In In submission, 2024

2023

  1. Pretrained Language Models as Visual Planners for Human Assistance
    Dhruvesh Patel, Hamid Eghbalzadeh, Nitin Kamra, Michael Louis Iuzzolino, Unnat Jain, and 1 more author
    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023

2022

  1. Modeling Label Space Interactions in Multi-label Classification using Box Embeddings
    Dhruvesh Patel, Pavitra Dangati, Jay-Yoon Lee, Michael Boratko, and Andrew McCallum
    In International Conference on Learning Representations, 2022
  2. Structured Energy Network As a Loss
    Jay Yoon Lee, Dhruvesh Patel, Purujit Goyal, Wenlong Zhao, Zhiyang Xu, and 1 more author
    In Advances in Neural Information Processing Systems, 2022

2020

  1. Weakly Supervised Medication Regimen Extraction from Medical Conversations
    Dhruvesh Patel, Sandeep Konam, and Sai Prabhakar
    In Proceedings of the 3rd Clinical Natural Language Processing Workshop, Nov 2020
  2. Representing Joint Hierarchies with Box Embeddings
    *Dhruvesh Patel, *Shib Sankar Dasgupta, Michael Boratko, Xiang Li, Luke Vilnis, and 1 more author
    In Automated Knowledge Base Construction (AKBC), Nov 2020