Nishant Mehta
University of Victoria
office: ECS 608 (Note: I don't check voicemail!)
n<lastname>@uvic.ca CV
Bio
My research focuses on developing and analyzing theoretically-principled methods for machine learning, which also may be called machine learning theory. In particular, I like to work on problems in online learning/sequential prediction (prediction with expert advice, multi-armed bandits, and reinforcement learning theory), statistical learning theory (empirical process theory, PAC-Bayesian methods), and transfer learning (especially lifelong learning and principled methods of transferring representations).
Prior to coming to Victoria, I previously was a postdoc at CWI in Amsterdam (with Peter Grünwald) and at the Australian National University in Canberra (with Bob Williamson). I did my PhD in the College of Computing at Georgia Tech, where I was advised by Alex Gray.
Prior to coming to Victoria, I previously was a postdoc at CWI in Amsterdam (with Peter Grünwald) and at the Australian National University in Canberra (with Bob Williamson). I did my PhD in the College of Computing at Georgia Tech, where I was advised by Alex Gray.
Papers
— also see my Google Scholar profile —
Beyond minimax rates in group distributionally robust optimization via a novel notion of sparsity
Quan Nguyen, Nishant Mehta, Cristóbal Guzmán
arXiv 2410.00690, 2024
On the price of exact truthfulness in incentive-compatible online learning with bandit feedback: a regret lower bound for WSU-UX
Ali Mortazavi, Junhao Lin, and Nishant Mehta
AISTATS 2024
Near-optimal per-action regret bounds for sleeping bandits
Quan Nguyen and Nishant Mehta
AISTATS 2024
An improved regret analysis for UCB-N and TS-N
Nishant Mehta
arXiv 2305.04093, 2023
Thresholded linear bandits
Nishant Mehta, Junpei Komiyama, Vamsi Potluru, Andrea Nguyen, Mica Grant-Hagen
AISTATS, 2023 - poster
Adversarial online multi-task reinforcement learning
Quan Nguyen and Nishant Mehta
ALT, 2023
long version
Best-case lower bounds in online learning
Cristóbal Guzmán, Nishant Mehta, Ali Mortazavi
NeurIPS, 2021 - poster
Optimal algorithms for private online learning in a stochastic environment
Bingshan Hu, Zhiming Huang, Nishant Mehta
arXiv 2102.07929, 2021
Fast rates for general unbounded loss functions: From ERM to generalized Bayes
Peter Grünwald and Nishant Mehta
JMLR, 2020
A Farewell to Arms: Sequential reward maximization on a budget with a giving up option
P Sharoff, Nishant Mehta, Ravi Ganti
AISTATS, 2020
Safe-Bayesian generalized linear regression
Rianne de Heide, Alisa Kirichenko, Peter Grünwald, Nishant Mehta
AISTATS, 2020
Dying experts: Efficient algorithms with optimal regret bounds
Hamid Shayestehmanesh, Sajjad Azami, Nishant Mehta
NeurIPS, 2019 - poster
Problem-dependent regret bounds for online learning with feedback graphs
Bingshan Hu, Nishant Mehta, Jianping Pan
UAI, 2019
Intelligent caching algorithms in heterogeneous wireless networks with uncertainty
Bingshan Hu, Yunjin Chen, Zhiming Huang, Nishant Mehta, Jianping Pan
IEEE International Conference on Distributed Computing Systems (ICDCS), 2019
Multi-observation regression
Rafael Frongillo, Nishant Mehta, Tom Morgan, and Bo Waggoner
AISTATS, 2019 - poster
A tight excess risk bound via a unified PAC-Bayesian-Rademacher-Shtarkov-MDL complexity
Peter Grünwald and Nishant Mehta
ALT, 2019 - slides
long version
Interpreting word-level hidden state behaviour of character-level LSTM language models
Avery Hiebert, Cole Peterson, Alona Fyshe, and Nishant Mehta
EMNLP, 2018 Workshop on Analyzing and Interpreting Neural Networks for NLP, 2018
Fast rates with high probability in exp-concave statistical learning
Nishant Mehta
AISTATS, 2017 - poster
Fast rates in statistical and online learning
Tim van Erven, Peter Grünwald, Nishant Mehta, Mark Reid, and Robert Williamson
JMLR, 2015 (Special issue in memory of Alexey Chervonenkis)
Generalized mixability via entropic duality
Mark Reid, Rafael Frongillo, Robert Williamson, and Nishant Mehta
COLT, 2015
From stochastic mixability to fast rates
Nishant Mehta and Robert Williamson
NIPS, 2014 (full oral presentation) - slides
long version
Sparsity-based generalization bounds for predictive sparse coding
Nishant Mehta and Alexander Gray
ICML, 2013
long version
MLPACK: A scalable C++ machine learning library
Ryan Curtin, James Cline, N.P. Slagle, William March, Parikshit Ram, Nishant Mehta, and Alexander Gray
JMLR, 2013
Minimax multi-task learning and a generalized loss-compositional paradigm for MTL
Nishant Mehta, Dongryeol Lee, and Alexander Gray
NIPS, 2012
version at NIPS 2012 Workshop on Multi-Trade-offs in Machine Learning
Computer detection approaches for the identification of phasic electromyographic (EMG) activity during human sleep
Jacqueline Fairley, George Georgoulas, Nishant Mehta, Alexander Gray, and Donald Bliwise
Biomedical Signal Processing and Control, 2012
Discriminative sparse coding for classification and regression
Nishant Mehta and Alexander Gray
The Learning Workshop (Snowbird), 2011 (oral presentation)
Generative and latent mean map kernels
Nishant Mehta and Alexander Gray
arXiv 1005.0188, 2010
Recognizing sign language from brain imaging
Nishant Mehta, Thad Starner, Melody Moore Jackson, Karolyn Babalola, and George Andrew James
International Conference on Pattern Recognition (ICPR), 2010
GVU Technical Report GIT-GVU-09-06, 2009
Optimal control strategies for an SSVEP-based brain-computer interface
Nishant Mehta, Sadhir Hussain, and Melody Moore Jackson
International Journal of Human-Computer Interaction, 2010
FuncICA for time series pattern discovery
Nishant Mehta and Alexander Gray
SIAM Data Mining, 2009 (nominated for best paper award; selected for oral presentation) - slides
Estimating neural signal dependence using kernels
Nishant Mehta, Alexander Gray, Thad Starner, and Melody Moore Jackson
NIPS 2008 Workshop on Statistical Analysis and Modeling of Response Dependencies in Neural Populations
Students
Quan Nguyen Manh, PhD studentAli Mortazavi, PhD student
Andrea Nguyen, MSc student
Mica Grant-Hagen, MSc student
Past students
Steve Scinocca, MSc
P Sharoff, MSc
Bingshan Hu, PhD, now a postdoc at UBC
Sajjad Azami, MSc
Hamid Shayestehmanesh, MSc
Yang Li, Undergrad, Starting a PhD at UWE Bristol
Junhao Lin, now a Master's student at Waterloo
Avery Hiebert, Undergrad, now a PhD student at Waterloo
Service
Action Editor for TMLR, 2022–Area Chair for COLT 2021–2024
Area Chair for ALT 2024
Area Chair for NeurIPS 2019, 2021–2024
Area Chair for ICML 2023
Area Chair for AAAI 2024, Senior PC member for AAAI 2021
I regularly review for the conferences AISTATS, COLT, NeurIPS/NIPS (prior to 2019), ICML (prior to 2021), ALT (sometimes)
I have reviewed for the journals JMLR, Machine Learning, JAIR, Artificial Intelligence, IEEE Transactions on Info Theory, PAMI, Bernoulli, Annales de l'Institut Henri Poincaré, JAIR, Artificial Intelligence, Canadian Journal of Statistics
Teaching
*Data Mining (CSC 503/SENG 474), Fall 2023. course webpageIntroduction to Online Learning (CSC 581A/482A), Spring 2023. course webpage
Data Mining (CSC 503/SENG 474), Spring 2023.
Algorithms and Data Structures (CSC 226), Fall 2022. course webpage
Data Mining (CSC 503/SENG 474), Spring 2022.
Machine Learning Theory (CSC 581A/482A), Fall 2021. course webpage
Algorithms and Data Structures (CSC 226), Fall 2021.
Data Mining (CSC 503/SENG 474), Spring 2021.
Algorithms and Data Structures (CSC 226), Spring 2021.
Data Mining (CSC 503/SENG 474), Summer 2020.
Data Mining (CSC 503/SENG 474), Spring 2020.
Machine Learning Theory (CSC 482A/581A), Fall 2019.
Algorithms and Data Structures (CSC 226), Fall 2019.
Machine Learning Theory (CSC 482A/581B), Spring 2019.
Algorithms and Data Structures (CSC 226), Fall 2018.
Machine Learning Theory (CSC 482A/581A), Spring 2018.
Algorithms and Data Structures (CSC 226), Fall 2017.
Information-Theoretic Learning (Leiden University, Spring 2016). Together with Peter Grünwald.
News
| July 2024 | I got tenure!!! |
| Jun-Aug, 2024 | I start a sabbatical visit to NYU, hosted by Junpei Komiyama |
| Feb-Apr, 2024 | I start a sabbatical visit to PUC in Santiago, Chile, hosted by Cristóbal Guzmán |
| Jan 2024 | Two papers accepted to AISTATS 2024! Quan's paper "Near-optimal per-action regret bounds for sleeping bandits" and Ali and Junhao's paper "On the price of exact truthfulness in incentive-compatible online learning with bandit feedback: a regret lower bound for WSU-UX" |
| Jan 19-20, 2023 | Pascale Gourdeau visits |
| Dec 2022 | Quan's paper "Adversarial online multi-task reinforcement learning" was accepted to ALT 2023! |
| Nov 3, 2022 | Jamie Morgenstern visits |
| Sep 9, 2022 | Sharan Vaswani visits |
| Sep 2021 | My Master's student, Steve Scinocca, successfully defended! |
| Aug 24-26, 2022 | Gautam Kamath visits and gives the first lecture for our new PIMS seminar series |
| Aug 6-16, 2022 | I visit Cristóbal Guzmán at PUC in Santiago, Chile |
| Jul 29-31, 2022 | I visit Vamsi Potluru and Mohsen Ghassemi at JP Morgan Chase AI Research in NYC |
| Mar 28-Apr 1, 2022 | I attend the Learning and Games program at Simons Institute in Berkeley |
| Feb 2022 | I'll be an Action Editor for the new journal TMLR! |
| Nov 2021 | Our PIMS seminar series on "The Mathematics of Ethical Decision-Making" funded! (together with Brandon Haworth, Valerie King, and Sowmya Somanath) |
| Nov 2021 | My Master's student, P Sharoff, successfully defended! |
| Sep 2021 | Ali's paper "Best-case lower bounds in online learning" (together with Cristóbal Guzmán) was accepted to NeurIPS 2021! |
| Sep 2021 | My first PhD student, Bingshan Hu, successfully defended!! Bingshan is headed to Amii as a Postdoctoral Fellow; mega-congrats Bingshan!!! |
| Jun 2021 | I was awarded a JP Morgan Chase Faculty Research Award |
| Feb 2021 | I'll be an Area Chair for COLT 2021 |
| Aug 2020 | My two Master's students, Sajjad Azami and Hamid Shayestehmanesh, both successfully defended! |
| Aug 2020 | I'll be a Senior PC member (Meta-Reviewer) for AAAI 2021 |
| Jan 2020 | Sharoff's paper "A Farewell to Arms" (together with Ravi Ganti) was accepted to AISTATS 2020! |
| Sep 2019 | Sajjad and Hamid's paper "Dying experts" was accepted to NeurIPS 2019! |
| May 2019 | Bingshan's paper "Problem-dependent regret bounds for online learning with feedback graphs" was accepted to UAI 2019! |
| Apr 2019 | I was awarded a New Frontiers in Research Grant (I’m Co-PI, joint with the amazing PI David Leitch) |
| Feb 2019 | I'll be an Area Chair for NeurIPS 2019 |
| Nov 2016 | I received an Outstanding Reviewer Award for NIPS 2016 (top 20 reviewers) |
Miscellanea
- Software
- MLPACK