Rajarshi Das (রাজর্ষি দাশ)

I am currently a researcher at AWS AI Labs in New York City, where I work on building retrieval and generative models over structured, unstructured, and multi-modal data. Previously, I was a postdoc at the wonderful H2lab at the University of Washington working with Prof. Hanna Hajishirzi. Before that, I completed my Ph.D. advised by Prof. Andrew McCallum as a part of the wonderful IESL lab at UMass Amherst. My Ph.D. thesis was on building neuro-symbolic models of reasoning over knowledge, primarily motivated by case-based reasoning.


My research interest lies in building semiparametric models of reasoning applied to structured (graphs, databases, tables), unstructured (text), and multimodal (images, UX widgets) data. I am interested in how new knowledge can be introduced (via nonparametric memories), used/manipulated (via parametric models), as well as synthesized/discovered (via reasoning).

Contact: dasrajar [at] amazon [dot] com. For JMLR related queries, please email at managing [at] jmlr [dot] org.

Selected Works

For a full list, check Google Scholar

  1. Bring Your Own KG: Self-Supervised Program Synthesis for Zero-Shot KGQA
    Dhruv Agarwal, Rajarshi Das, Sopan Khosla, Rashmi Gangadharaiah NAACL 2024
    [code], [Slides]
  2. When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories
    Alex Mallen, Akari Asai, Victor Zhong, Rajarshi Das, Daniel Khashabi, Hannaneh Hajishirzi ACL 2023
    [code]
  3. Nonparametric Contextual Reasoning for Question Answering over Large Knowledge Bases
    Rajarshi Das Ph.D. Thesis 2022
  4. Case-based Reasoning for Natural Language Queries over Knowledge Bases
    Rajarshi Das, Manzil Zaheer, Dung Thai, Ameya Godbole, Ethan Perez, Jay-Yoon Lee, Lizhen Tan, Lazaros Polymenakos, Andrew McCallum EMNLP 2021
  5. A Simple Approach to Case-Based Reasoning in Knowledge Bases
    Rajarshi Das, Ameya Godbole, Shehzaad Dhuliawala, Manzil Zaheer, Andrew McCallum AKBC 2020 Best Paper Runner-up
    [code], [Talk]
  6. Multi-step Retriever-Reader Interaction for Scalable Open-domain Question Answering
    Rajarshi Das, Shehzaad Dhuliawala, Manzil Zaheer, Andrew McCallum ICLR 2019
    [code]
  7. Go for a Walk and Arrive at the Answer -- Reasoning over Paths in Knowledge Bases using Reinforcement Learning
    Rajarshi Das, Shehzaad Dhuliawala, Manzil Zaheer, Luke Vilnis, Ishan Durugkar, Akshay Krishnamurthy, Alex Smola, Andrew McCallum ICLR 2018
    [code]

Mentoring / Interns

I had the pleasure of mentoring and working with several graduate and undergraduate students during my PhD. I am grateful that I get to continue to do that in industry by hosting interns.


Invited Talks

Apr 2024: Talk at Georgia Tech on Semiparametric Reasoning over Structured Data
May 2022: Talk at CMU
February 2022: Talk at Stanford NLP seminar series on Nonparametric Contextual Reasoning for Question Answering over Knowledge Bases
June 2021: Talk at University of Washington (H2Lab)

Service

Along with Tegan, I serve as a managing editor for Journal of Machine Learning Research (JMLR)
I had a great time co-organizing the weekly Machine Learning and Friends Lunch for 3 years. Please consider giving a talk!
I have co-organized the following workshops I review for almost all of major NLP (ACL, EMNLP, NAACL, EACL) and ML conferences (Neurips, ICML, ICLR) and journal (TMLR) every year.