Nikunj Saunshi

Department of Computer Science |

About

I am a PhD candidate in the Computer Science department at Princeton University, fortunate to be advised by Prof. Sanjeev Arora. My interests lie in understanding the success of modern machine learning algorithms and improving/designing algorithms that are grounded in mathematical principles. I am broadly interested in Theoretical Machine Learning and Natural Language Processing with recent focus on theory for self-supervised learning, representation learning and meta-learning.
I received a Masters of Science in Engineering in Computer Science at Princeton prior to this. Before joining Princeton University, I worked as an Associate Engineer at Samsung Electronics, South Korea in the R&D center. I completed my B.Tech. with Honors in Computer Science and Engineering and Minor in Mathematics from the Indian Institute of Technology Bombay.
Research

(#) denotes alphabetical order
Preprints

- Predicting What You Already Know Helps: Provable Self-Supervised Learning

(#) Jason D. Lee, Qi Lei, Nikunj Saunshi, Jiacheng Zhuo

*Preprint*

Publications

- A Mathematical Exploration of Why Language Models Help Solve Downstream Tasks

Nikunj Saunshi, Sadhika Malladi, Sanjeev Arora

*To Appear in ICLR 2021*

[Talk] - A Sample Complexity Separation between Non-Convex and Convex Meta-Learning

Nikunj Saunshi, Yi Zhang, Mikhail Khodak, Sanjeev Arora*ICML 2020*

[Talk] - Provable Representation Learning for Imitation Learning via Bi-level Optimization

(#) Sanjeev Arora, Simon S. Du, Sham Kakade, Yuping Luo, Nikunj Saunshi*ICML 2020* - A Theoretical Analysis of Contrastive Unsupervised Representation Learning

(#) Sanjeev Arora, Hrishikesh Khandeparkar, Mikhail Khodak, Orestis Plevrakis, Nikunj Saunshi*ICML 2019*

[Blog] [Poster] [Talk] - A La Carte Embedding: Cheap but Effective Induction of Semantic Feature Vectors

Mikhail Khodak*, Nikunj Saunshi*, Yingyu Liang, Tengyu Ma, Brandon Stewart and Sanjeev Arora*ACL 2018*

[Blog] [Slides] [Talk] - A Compressed Sensing View of Unsupervised Text Embeddings, Bag-of-n-Grams, and LSTMs

(#) Sanjeev Arora, Mikhail Khodak, Nikunj Saunshi, Kiran Vodrahalli*ICLR 2018*

[Blog] [Poster] - A Large Self-Annotated Corpus for Sarcasm

Mikhail Khodak, Nikunj Saunshi, Kiran Vodrahalli

*LREC 2018*

Teaching

Princeton University

- Spring 18: Assistant in Instruction for COS 340 - Reasoning about Computation
- Fall 17: Assistant in Instruction for COS 324 - Introduction to Machine Learning
- Spring 17: Assistant in Instruction for COS 340 - Reasoning about Computation
- Fall 16: Assistant in Instruction for COS 340 - Reasoning about Computation

Indian Institute of Technology Bombay

- Spring 14: Teaching Assistant for MA 106 - Linear Algebra
- Fall 12: Teaching Assistant for CS 207 - Discrete Structures
- Spring 12: Teaching Assistant for MA 108 - Differential Equations
- Fall 11: Teaching Assistant for MA 105 - Calculus

Contact

Department of Computer SciencePrinceton University

35 Olden Street

Princeton NJ - 08540

Email: nsaunshi{at}cs{dot}<univ>{dot}edu