Scholar
Shivam Garg
Google Scholar ID: lWzkIg4AAAAJ
Senior Researcher, Microsoft Research
Citations & Impact
Citations
872
H-index
9
i10-index
9
Publications
15
Co-authors
14
list available
Contact
Publications
Resume (English only)
Academic Achievements
- - Publications:
- - On the Statistical Complexity of Sample Amplification, arXiv 2022
- - What Can Transformers Learn In-Context? A Case Study of Simple Function Classes, NeurIPS 2022 (oral presentation)
- - How and When Random Feedback Works: A Case Study of Low-Rank Matrix Factorization, AISTATS 2022
- - Distributed Algorithms from Arboreal Ants for the Shortest Path Problem, PNAS 2022 (to appear), ITCS 2021
- - Reward Identification in Inverse Reinforcement Learning, ICML 2021
- - Sample Amplification: Increasing Dataset Size even when Learning is Impossible, ICML 2020, NeurIPS Workshop on Machine Learning with Guarantees 2019 (oral presentation)
- - A Spectral View of Adversarially Robust Features, NeurIPS 2018 (spotlight presentation)
- - Raising the Bar for Vertex Cover: Fixed-parameter Tractability Above a Higher Guarantee, SODA 2016
Research Experience
- - Member of the Machine Learning Group and Theory Group at Stanford.
- - Co-founder and organizer of Algorithms and Friends, an initiative to increase interaction between the Theory Group and other Stanford researchers.
- - Worked at Microsoft Research India for a year, collaborating with Deeparnab Chakrabarty and Ravishankar Krishnaswamy.
Education
- - Stanford University, PhD (in progress)
- - Advisor: Greg Valiant
- - Time: Present
- - Field: Machine Learning
- - Indian Institute of Technology Bombay, Bachelor's Degree
- - Time: Before Stanford
- - Microsoft Research India, worked for one year
- - Collaborators: Deeparnab Chakrabarty and Ravishankar Krishnaswamy
Background
- - Research Interests: Understanding the foundations of intelligence, both artificial and natural.
- - Fields: Machine learning, statistics, algorithms, and biology.
- - Brief Introduction: A PhD student at Stanford, focusing on improving the algorithmic and reasoning capabilities of modern machine learning systems, efficiently and robustly extracting information from data, and investigating how intelligence emerges in natural systems using a computational lens.
Miscellany
- - Organizes Algorithms and Friends, helping to solve algorithmic questions in applied research and organizing seminars where researchers from diverse backgrounds share their work.
