Jayneel Parekh
Postdoc @ ISIR, Sorbonne Université
Hi there! I am currently a Postdoc in MLIA team at ISIR, Sorbonne Univerity with Prof. Matthieu Cord, working on understanding and ultimately enhancing large multimodal models. Previously, I was a PhD student at LTCI, Telecom Paris under the supervision of Prof. Florence d’Alche and Prof. Pavlo Mozharovskyi, on the topic of interpretability of neural networks, applied on image or audio data. I completed my undergrad studies in Electrical Engineering at IIT Bombay with focus on Signal Processing, where I worked with Prof. Preeti Rao and Prof. Yi-Hsuan Yang on Speech-to-Singing conversion.
Broadly, I am interested by machine learning applications in audio/visual/multimodal domains. Barring research, I enjoy chess, squash, calisthenics, or following sports (football, boxing, cricket etc.).
news
| Jan 24, 2025 | Paper “Restyling Unsupervised Concept Based Interpretable Networks with Generative Models” accepted in ICLR 2025! |
|---|---|
| Sep 28, 2024 | Paper “A Concept-Based Explainability Framework for Large Multimodal Models” accepted in NeurIPS 2024! |
| Jun 25, 2024 | L2I paper awarded 2nd prize for STIC Best Scientific Contribution 2023 |
| Dec 26, 2023 | Paper “Tackling interpretability in audio classification networks with non-negative matrix factorization” accepted in IEEE/ACM TASLP journal! |
| Dec 1, 2023 | Top Reviewer at NeurIPS 2023! |
| Sep 5, 2023 | Talk at Deezer Research Paris. |
| Jul 26, 2023 | Organizing a ‘Robustness, Fairness and Explainability’ workshop at ELLIS Unconference 2023 Paris with Quentin Bouniot. |
| Jul 7, 2023 | Successfully defended my thesis! |
| Dec 4, 2022 | Presented L2I at NeurIPS@Paris and NeurIPS 2022 in-person conference (New Orleans). Both amazing events! |
| Sep 15, 2022 | Paper: “Listen to Interpret: Post-hoc Interpretability for Audio Networks with NMF” (L2I) accepted at NeurIPS 2022! |
| Aug 15, 2022 | Attended Oxford Machine Learning Summer School (ML x Finance track, in-person). |
| Jun 14, 2022 | Attended + Presented FLINT at Franco-German workshop on AI at INRIA Rocquencourt (in “Trustworthy and Explainable Artificial Intelligence” session). |
| May 10, 2022 | Talk at DataCraft Paris |
| Jan 5, 2022 | Talk for FLINT at IBM Research, Saclay. Presented by my advisor (I was down with fever!) |
| Nov 5, 2021 | Talk (for FLINT) at IDEMIA Paris. |
| Oct 12, 2021 | Paper: “A Framework to Learn with Interpretation” (FLINT) accepted at NeurIPS 2021! |
| Aug 23, 2021 | Attended MLSS-2021 (Virtually) |