Jyothir S V

I am a Jr. Research Scientist at the Center for Data Science at New York University, where I am advised by Yann LeCun and Kyunghyun Cho.

My research interests lie at the intersection of reinforcement learning, self-supervised learning, humanoid control and enhancing mathematical reasoning capabilities within large language models.

I completed my undergraduate studies at the Indian Institute of Technology (IIT) Mandi located in Himalayas, where I conducted research under the mentorship of Aditya Nigam.

Please feel free to contact me: jyothir -at- nyu.edu

Email  /  GitHub  /  Google Scholar  /  LinkedIn

profile photo

Research

project image

Hierarchical World Models as Visual Whole-Body Humanoid Controllers


Nicklas Hansen, Jyothir S V, Vlad Sobal, Yann LeCun, Xiaolong Wang, Hao Su
ICLR Submission 2025, CoRL Workshop on Whole-body Control and Bimanual Manipulation 2024
arxiv / code /

In this paper, we introduce a hierarchical world model for whole-body humanoid control that learns dynamics from human motion captured data. We also propose a novel task suite for evaluation and highlight the effectiveness of our model through various experiments.

project image

Recite, Reconstruct, Recollect: Memorization in LMs as a Multifaceted Phenomenon


USVSN Sai Prashanth*,..., Jyothir S V*,..., Katherine Lee, Naomi Saphra
ICLR Submission 2025, EMNLP Finding 2024
arxiv / code /

In this work, we propose a taxonomy for memorization in language models, breaking it down into recitation, reconstruction, and recollection. We demonstrate the taxonomy’s utility by constructing predictive models for memorization and analyzing dependencies that differ across categories and features.

project image

Gradient-based Planning with World Models


Jyothir S V, Siddhartha Jalagam, Yann LeCun, Vlad Sobal
ICLR Generative Models for Decision Making Workshop 2024
arxiv / code /

Most model predictive control (MPC) algorithms designed for visual world models have traditionally explored gradient-free population-based optimization methods, such as Cross Entropy and Model Predictive Path Integral (MPPI) for planning. We present an exploration of a gradient-based alternative that fully leverages the differentiability of the world model.

project image

A Massively Multi-System MultiReference Data Set for Dialog Metric Evaluation


Huda Khayrallah, Zuhaib Akhtar, Edward Cohen, Jyothir S V , João Sedoc
EMNLP Findings 2023
paper /

Automatic metrics for dialogue evaluation should be robust proxies for human judgments; however, the verification of robustness is currently far from satisfactory. To quantify the robustness correlation and understand what is necessary in a test set, we create and release an 8-reference dialog datase. We then train 1750 systems and evaluate them and publicly available large models on our novel test set and the DailyDialog dataset.

project image

Joint Embedding Predictive Architectures Focus on Slow Features


Vlad Sobal, Jyothir S V, Siddhartha Jalagam, Nicholas Carion, Kyunghyun Cho, Yann Lecun
NeurIPS SSL Workshop 2022
arxiv / code /

In this work, we analyze performance of JEPA trained with VICReg and SimCLR objectives in the fully offline setting without access to rewards, and compare the results to the performance of the generative architecture.





Design and source code from Jon Barron's website