Debjit Paul

Postdoc at EPFL

DSC0818.jpg
Hi there! I'm a Postdoctoral Researcher working with Boi Faltings, Robert West, and Antoine Bosselut at EPFL. Before EPFL, I completed my Ph.D. under the supervision of Anette Frank in the Department of Computational Linguistics at Heidelberg University and also a part of Research Training Group AIPHES.
My overarching research objective is to enhance the reasoning capabilities of NLP systems by advancing their ability to understand the complex interplay between language, knowledge representation, and cognitive reasoning processes. To accomplish this goal, I have focused my investigations in three areas:
  • Reasoning with knowledge representations - NLP systems must be grounded by knowledge dynamics to generalize to unseen situations robustly. My research has focused on designing computational models capable of reasoning over knowledge.
  • Learning to generate explanations for reasoning - When a model explains its decision-making steps or logical steps it took to reach that conclusion, it enhances transparency and helps build trust in the system. My research has focused on designing computational models that generate explanations before making the final decisions.
  • Aligning Models with Human or AI Feedback - Text generation models are observed to display undesired and inconsistent behaviours, such as hallucination and unfaithful reasoning. My research has focused on designing methods to rectify the undesired behaviours of text generation models through interaction.

News

Feb 21, 2024 Check out our new preprint paper Making Reasoning Matter: Measuring and Improving Faithfulness of Chain-of-Thought Reasoning
Jan 20, 2024 REFINER got accepted at EACL main conference 2024. See you in Malta :airplane:
Oct 9, 2023 Two papers got accepted at EMNLP 2023. See you in Singapore :airplane:
Aug 13, 2023 I am on the academic and industry job market. If you think my background can be a good fit, I’d be happy to discuss!
Aug 2, 2023 Check out our new preprint paper Flows: Building Blocks of Reasoning and Collaborating AI

latest posts

selected publications

2024

  1. Frodo.png
    Making Reasoning Matter: Measuring and Improving Faithfulness of Chain-of-Thought Reasoning
    Debjit PaulRobert WestAntoine Bosselut, and Boi Faltings
    In , Feb 2024
  2. Refiner.png
    REFINER: Reasoning Feedback on Intermediate Representations
    Debjit Paul, Mete Ismayilzada, Maxime Peyrard, Beatriz Borges, Antoine BosselutRobert West, and Boi Faltings
    In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), Mar 2024

2023

  1. CROW.png
    CRoW: Benchmarking Commonsense Reasoning in Real-World Tasks
    Mete Ismayilzada, Debjit Paul, Syrielle Montariol, Mor Geva, and Antoine Bosselut
    In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP), Dec 2023