Sai Ganesh Nagarajan

Welcome

Bio

I am a postdoc and a thrust leader of the Robust and Explainable Learning Group at Zuse Institute Berlin which is under Prof. Sebastian Pokutta's group . Prior to this I was a postdoc at EPFL under Prof. Ola Svensson. Before that, I was fortunate to graduate from Singapore University of Technology and Design (SUTD), generously supported by the SUTD President's Graduate PhD Fellowship (PGF) and during this time I had the excellent guidance of Georgios Piliouras (SUTD) and Ioannis Panageas (UC Irvine)!

Research Interests

I am interesting in understanding the foundations of multiagent AI. I study the efficacy of learning dynamics and efficiency of minmax optimization. Furthermore, I am interested in applications of game theory/minmax optimization to problems in AI safety.

I was a PC member of AAAI 2025 and AAMAS 2023. In general I serve as a reviewer for ICML,NEURIPS, AISTATS, ICLR, TMLR. In addition, I have reviewed papers for FOCS, Algorithmica, EC, WINE, SAGT.

I am on the job market this year, looking for academic/relevant industry research positions! Here is my CV . Please feel free to reach out to discuss about my research!

Selected Publications

Alexandros Hollender, Gilbert Maystre, Sai Ganesh Nagarajan. The Complexity of Two-Team Polymatrix Games with Independent Adversaries. [arxiv]. (ICLR 2025, Oral Presentation (Top 1.8%)).(Alphabetical Order)
Etienne Bamas, Sai Ganesh Nagarajan, Ola Svensson. Analyzing $D^\alpha$ seeding for $k$-means. [paper][code]. (ICML 2024). (Alphabetical Order)
Marina Drygala, Sai Ganesh Nagarajan, Ola Svensson. Online Algorithms with Costly Predictions. [paper]. (AISTATS 2023). (Alphabetical Order)
Qi Lie, Sai Ganesh Nagarajan, Ioannis Panageas and Xiao Wang. Last iterate convergence in no-regret learning: constrained min-max optimization for convex-concave landscapes. [arxiv]. (AISTATS 2021). (Alphabetical Order)
Sai Ganesh Nagarajan, David Balduzzi and Georgios Piliouras. From Chaos to Order: Symmetry and Conservation Laws in Game Dynamics. [paper] [code] (ICML 2020).

Contact details:

nagarajanATzibDOTde

sganeshDOT22ATgmailDOTcom

News

I am excited to travel to Asia, giving talks at ICLR, Singapore and DOML @RIMS, Kyoto!
"The Good, the Bad and the Ugly: Watermarks, Transferable Attacks and Adversarial Defenses" is accepted at ICLR 2025 Workshop on GenAI Watermarking.
Giving a contributed talk on "The Complexity of Two-Team Polymatrix Games with Independent Adversaries" at the 15th Days on Computational Game Theory, Tilburg.
"The Complexity of Two-Team Polymatrix Games with Independent Adversaries" is accepted at ICLR 2025 as an Oral Presentation!
"S-CFE: Simple Counterfactual Explanations" is accepted at AISTATS 2025!
Recently gave talks on "The Good, the Bad and the Ugly: Watermarks, Transferable Attacks and Adversarial Defenses" at the National University of Singapore , Nanyang Technological University and the Singapore University of Technology and Design.
New paper on "The Complexity of Two-Team Polymatrix Games with Independent Adversaries" out! [arxiv].
I am thrilled to be teaching a course on Modern Algorithms for Multiagent Learning at TU Berlin during the Winter Semester 2024/25! [course].
Gave a talk on "Analyzing $D^\alpha$ seeding for k-means" at Khoury College of Computing, Northeastern University, Boston.
Excited to serve as a PC member for AAAI 2025!
"Unified Taxonomy of AI Safety: Watermarks, Adversarial Defenses and Transferable Attacks" is accepted at ICML 2024 Workshop on Theoretical Foundations of Foundation Models [paper].
"Analyzing $D^\alpha$ seeding for k-means" is accepted at ICML 2024 [paper] .
I started a postdoc at Zuse Institute Berlin in Prof.Sebastian Pokutta's group from Nov 2023 onwards!
PhD student Marina Drygala is presenting a poster of our work "Online Algorithms with Costly Predictions" in the Workshop on Learning-augmented Algorithms: Theory and Applications at SIGMETRICS 2023.
"Online Algorithms with Costly Predictions" is accepted at AISTATS 2023.
"Mean estimation of truncated mixtures of two Gaussians: A gradient based approach" is accepted at AAAI 2023.
I am excited to serve as a PC member for AAMAS 2023.
I am grateful to be recognized as a Top Reviewer for NeurIPS 2022.
In July 2022, I was a TA for the course on the Interplay between Optimization and Sampling in High Dimensions, by Santosh Vempala, as a part of the EPFL Summer School on Modern Trends in Combinatorial Optimization.
Two papers accepted to AISTATS 2021.
I defended my Phd thesis titled "Machine Learning via Dynamical Systems: Applications to Deep Learning, Game Theory and Optimization" successfully on January 20th 2021!
In 2021, I will be joining as a postdoc, hosted by Prof. Ola Svensson at the EPFL CS theory group!
Virtual Talk on "On the analysis of EM for truncated mixtures of two Gaussians" at ACAC 2020.
Virtual Talk on "From Chaos to Order: Symmetry and Conservation Laws in Game Dynamics" at ICML 2020.
Two papers accepted to ICML 2020.
Talk on "On the analysis of EM for truncated mixtures of two Gaussians" at ALT 2020, San Diego.
Presented my work "On the analysis of EM for truncated mixtures of two Gaussians" as a poster at MIT Foundations of Data Science workshop: Learning under Complex Structure from Jan 27th-Jan 29th 2020.
Recent work on "Depth-Width Trade-offs for ReLU Networks via Sharkovsky's Theorem" has been accepted as a spotlight presentation at ICLR 2020.
My work with Ioannis Panageas "On the analysis of EM for truncated mixtures of two Gaussians" is accepted for presentation in ALT 2020.
Awarded the prestigious Merit Award in the Minister for National Development's R & D Awards 2019 for my contribution in the Urban Microclimate Multiphysics Integrated Simulation Tool (UM-MIST) project done at A*STAR.
Contributed Talk in the Learning in the Presence of Strategic Behavior Workshop at Economics and Computation 2019, FCRC, Phoenix.
Attended the Deep Learning Bootcamp in the Simons Institute for the Theory of Computing at UC Berkeley from May 26th 2019- May 31st 2019.
Teaching Assistant for Simulation,Modelling and Analysis from Sept 2018-Dec 2018 and Jan 2019-May 2019 (TA led course).
Invited to attend the summer school on "Optimal Transport meets Economic Theory" at the Hausdorff Center of Mathematics, University of Bonn from Jul 23rd-28th 2018.
Talk on "Three Body Problems in Evolutionary Game Theory: Convergence, Periodicity and Limit cycles" at AAMAS 2018, Stockholm.

Google Sites

Report abuse