Hi! My name is Tatjana (pronounced /tatiana/), currently a Visiting Professor in the Department of Electronics, Information and Bioengineering (DEIB) at Politecnico di Milano (Polimi), where I collaborate with Nicola Gatti.
I earned my Ph.D. from EPFL, and Idiap, advised by François Fleuret. During my doctoral studies, I completed internships at (i) Mila where I was supervised by Yoshua Bengio and Simon Lacoste-Julien, as well as at (ii) DeepMind mentored by Irina Higgins. Following my Ph.D., I worked as a Postdoctoral Research Scientist at the Machine Learning and Optimization (MLO) lab at EPFL, advised by Martin Jaggi. Later, I joined the Department of Electrical Engineering and Computer Science (EECS) at UC Berkeley as a Postdoctoral Researcher, working with Michael Jordan.
🏅 Thanks to the Swiss National Science Foundation for supporting my research, through the Early.Postdoc.Mobility (in 2021) and the Postdoc.Mobility (in 2023) fellowships.
Research Interests
My research is at the intersection of game theory and machine learning (or commonly referred to as AI). My work aims to better understand the learning dynamics of multi-player games, where players are neural nets, and develop improved learnig methods. Beyond this, I’m also interested in multi-agent reinforcement learning, unsupervised learning, generative modeling, generalization, and robustness.
News:
Sept., 2024: Very excited to be working with Nicola Gatti, Alberto Marchesi, Matteo Castiglioni, and the remaining game theory and reinforcement learning team at Polimi and Milano! July, 2024: I will be at the ISMP conference and talk about HRDEs for VIs (slides coming soon). See you in Montrèal! PS - if you're attending ICML, don't miss out WiML's events! Jan, 2024: You have two- or multi-player problem with constraints? use our ACVI method -> accepted at ICLR 2024. See you in Vienna! October 23, 2023: I'll join CISPA as a tenure-track faculty, and will give a course on Games in Machine Learning (winter semester) at Saarland University in Germany -- see the course materials! If you are interested in joining my group use this form to apply! August, 2023: I'll be at the EUROPT workshop in Budapest; let me know if you'll be there! mid June, 2023: I will be at Jordan Symposium in France, let me know if you're there! May 31 - June 3, 2023: I will be at SIAM OP23 in Seattle, and talk about solving VIs with constraints: slides May 1 - May 5, 2023: I will attend ICLR in person and talk about how to solve games with constraints
Selected Research
-
(preprint)
Learning Variational Inequalities from Data: Fast Generalization Rates under Strong Monotonicity
Eric Zhao, Tatjana Chavdarova, Michael I. Jordan
ArXiv, 2024 • Paper • BibTex@article{zhao2024learningvariationalinequalities, title = {Learning Variational Inequalities from Data: Fast Generalization Rates under Strong Monotonicity}, author = {Eric Zhao and Tatjana Chavdarova and Michael I. Jordan}, journal = {ArXiv:2410.20649}, year = {2024}, }
-
(workshop)
Variational Inequality Methods for Multi-Agent Reinforcement Learning:
Performance and Stability Gains
Baraah A. M. Sidahmed, Tatjana Chavdarova
OWA-2024 Workshop at NeurIPS, 2024 • Paper • BibTex@article{sidahmed2024vimarl, title = {Variational Inequality Methods for Multi-Agent Reinforcement Learning: Performance and Stability Gains}, author = {Baraah A. M. Sidahmed and Tatjana Chavdarova}, journal = {ArXiv:2410.07976}, year = {2024}, }
-
(workshop)
On the Hypomonotone Class of Variational Inequalities
Khaled Alomars, Tatjana Chavdarova
OPT-2024 Workshop at NeurIPS, 2024 • Paper • BibTex@article{alomar2024hypomonotone, title = {On the Hypomonotone Class of Variational Inequalities}, author = {Khaled Alomar and Tatjana Chavdarova}, journal = {ArXiv:2410.09182}, year = {2024}, }
-
(workshop)
Decoupled Stochastic Gradient Descent for N-Player Games
Ali Zindari, Parham Yazdkhasti, Tatjana Chavdarova, Sebastian U. Stich
ARLET Workshop at ICML, 2024 • Paper • BibTex@article{zindari2024decoupled, title = {Decoupled Stochastic Gradient Descent for N-Player Games}, author = {Ali Zindari and Parham Yazdkhasti and Tatjana Chavdarova and Sebastian U Stich}, journal = {ICML 2024 Workshop: Aligning Reinforcement Learning Experimentalists and Theorists}, year = {2024}, }
-
A Primal-Dual Approach to Solving Variational Inequalities with General Constraints
Tatjana Chavdarova*, Tong Yang*, Matteo Pagliardini, Michael I. Jordan
ICLR, 2024 • Paper • Poster • Code • BibTex@inproceedings{chavdarova2024acvi, title = {A Primal-dual Approach for Solving Variational Inequalities with General-form Constraints}, author = {Chavdarova, Tatjana and Yang, Tong and Pagliardini, Matteo and Jordan, Michael I.}, booktitle = {ICLR}, year = {2024}, }
-
Last-Iterate Convergence of Saddle Point Optimizers via High-Resolution Differential Equations
Tatjana Chavdarova*, Michael I. Jordan*, Manolis Zampetakis*
MTA journal, 2023 & early version at OPT NeurIPS workshop, 2021 • Paper • ArXiv • Poster • BibTex@article{chavdarova2023hrdes, title = {Last-Iterate Convergence of Saddle Point Optimizers via High-Resolution Differential Equations}, author = {Tatjana Chavdarova and Michael I. Jordan and Manolis Zampetakis}, journal = {Minimax Theory and its Applications}, year = {2023}, }
-
Solving Constrained Variational Inequalities via a First-order Interior Point-based Method—Notable Paper (top 25%)
Tong Yang*, Michael I. Jordan*, Tatjana Chavdarova*
ICLR, 2023 • Paper • Slides • Poster • Code • BibTex@inproceedings{yang2023acvi, title = {Solving Constrained Variational Inequalities via a First-order Interior Point-based Method}, author = {Tong Yang and Michael I. Jordan and Tatjana Chavdarova}, booktitle = {ICLR}, url={https://openreview.net/forum?id=RQY2AXFMRiu}, year = {2023}, }
-
(workshop)
Continuous-time Analysis for Variational Inequalities: An Overview and Desiderata
Tatjana Chavdarova*, Ya-Ping Hsieh*, Michael I. Jordan*
Continuous time methods for ML, ICML workshop, 2022 • ArXiv • Poster • BibTex@article{chavdarova2022contVIs, title = {Continuous-time Analysis for Variational Inequalities: An Overview and Desiderata}, author = {Tatjana Chavdarova and Ya-Ping Hsieh and Michael I. Jordan}, booktitle= {ICML Workshop on Continuous time methods for Machine Learning}, year = {2022}, }
-
(workshop)
The Peril of Popular Deep Learning Uncertainty Estimation Methods
Yehao Liu*, Matteo Pagliardini*, Tatjana Chavdarova, Sebastian U. Stich
BDL NeurIPS workshop, 2021 • Paper • Code • Poster • BibTex@article{liu2021peril, title = {The Peril of Popular Deep Learning Uncertainty Estimation Methods}, author = {Yehao Liu and Matteo Pagliardini and Tatjana Chavdarova and Sebastian U. Stich}, journal = {NeurIPS Workshop on Bayesian Deep Learning}, year = {2021}, }
-
(workshop)
Improved Adversarial Robustness via Uncertainty Targeted Attacks
Gilberto Manunza*, Matteo Pagliardini*, Martin Jaggi, Tatjana Chavdarova
UDL ICML workshop, 2021 • Paper • Poster • BibTex@article{ManunzaPagliardini2021, title = {Improved Adversarial Robustness via Uncertainty Targeted Attacks}, author = {Gilberto Manunza and Matteo Pagliardini and Martin Jaggi and Tatjana Chavdarova}, journal = {ICML Workshop on Uncertainty and Robustness in Deep Learning}, year = {2021}, url = {http://www.gatsby.ucl.ac.uk/~balaji/udl2021/accepted-papers/UDL2021-paper-096.pdf} }
-
Semantic Perturbations with Normalizing Flows for Improved Generalization
Oguz K Yüksel, Sebastian U. Stich, Martin Jaggi, Tatjana Chavdarova
ICCV 2021 • Paper • Code • Video • BibTex@InProceedings{Yuksel_2021_ICCV, author = {Y\"uksel, Oguz Kaan and Stich, Sebastian U. and Jaggi, Martin and Chavdarova, Tatjana}, title = {Semantic Perturbations With Normalizing Flows for Improved Generalization}, booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)}, month = {October}, year = {2021}, pages = {6619-6629} }
-
Taming GANs with Lookahead-Minmax
Tatjana Chavdarova*, Matteo Pagliardini*, Sebastian U. Stich, François Fleuret, Martin Jaggi
ICLR 2021 •Paper •Code •Video •Poster •BibTex@inproceedings{chavdarova2021lagan, author = {Tatjana Chavdarova and Matteo Pagliardini and Sebastian U. Stich and Fran{\c{c}}ois Fleuret and Martin Jaggi}, title = {Taming GANs with Lookahead-Minmax}, booktitle = {{ICLR}}, publisher = {OpenReview.net}, year = {2021} }
-
(PhD thesis)
Deep Generative Models and Applications
Tatjana Chavdarova, Jury: F. Fleuret, P. Frossard, L. Denoyer, S. Lacoste-Julien, M. Jaggi.
EPFL, July 2020 •Thesis •BibTex@article{Chavdarova:278463, title = {Deep Generative Models and Applications}, author = {Chavdarova, Tatjana}, institution = {IEL}, publisher = {EPFL}, address = {Lausanne}, pages = {169}, year = {2020}, url = {http://infoscience.epfl.ch/record/278463}, doi = {10.5075/epfl-thesis-10257}, }
-
Reducing Noise in GAN Training with Variance Reduced Extragradient
Tatjana Chavdarova*, Gauthier Gidel*, François Fleuret, Simon Lacoste-Julien
NeurIPS 2019 •Paper •Code •Slides •Poster •Video •BibTex@inproceedings{chavdarova2019, Author = {Tatjana Chavdarova and Gauthier Gidel and François Fleuret and Simon Lacoste-Julien}, Title = {Reducing Noise in {GAN} Training with Variance Reduced Extragradient}, Booktitle = {{Advances in Neural Information Processing Systems (NeurIPS)}}, Year = {2019}, volume = {32}, publisher = {Curran Associates, Inc.}, }
‡Selected as oral at MBRL workshop @ICML, and at Montreal AI Symposium '19. -
SGAN: An Alternative Training of Generative Adversarial Networks
Tatjana Chavdarova, François Fleuret
CVPR 2018 •Paper •Slides •Poster •BibTex@inproceedings{chavdarova-fleuret-2018, author = {Chavdarova, T. and Fleuret, F.}, title = {{SGAN}: An Alternative Training of Generative Adversarial Networks}, booktitle = {CVPR}, year = {2018}, }
-
WILDTRACK: A Multi-camera HD Dataset for Dense Unscripted Pedestrian Detection
T. Chavdarova, P. Baqué, S. Bouquet, A. Maksai, T. Bagautdinov, L. Lettry, P. Fua, L. Van Gool, François Fleuret
CVPR 2018 •Paper •Dataset •Code •Poster •BibTex@inproceedings{chavdarova-et-al-2018, author = {Chavdarova, T. and Baqué, P. and Bouquet, S. and Maksai, A. and Jose, C. and Bagautdinov, T. and Lettry, L. and Fua, P. and Van Gool, L. and Fleuret, F.}, title = {{WILDTRACK}: A Multi-camera {HD} Dataset for Dense Unscripted Pedestrian Detection}, booktitle = {CVPR}, year = {2018}, pages = {5030-5039}, }
* Equal contributions.
Selected Talks
- A First Order Primal-Dual Method for Solving Constrained Variational Inequalities
20th EUROpt Workshop, Budapest, Hungary, August 2023. • slides - Learning Dynamics in Multi-Player Games
Robotics and Perception Group of Prof. Davide Scaramuzza, University of Zurich, Switzerland; July 2023. • slides - A First Order Primal-Dual Method for Solving Constrained Variational Inequalities
SIAM OP23, Seattle, WA; May 31, 2023. • abstract • slides ‡ My coauthor Manolis Zampetakis presented our work on HRDEs, June 1, 2023. • slides - Beyond Standard Minimization: Solving for Equilibria in Constrained Variational Inequalities
Symposium on Frontiers of Machine Learning and Artificial Intelligence, Viterbi School of Engineering, University of Southern California, Nov. 2022 • slides - Optimization of differentiable games
NSA Lab, Johns Hopkins University, April 2022 • slides - Continious-time Tools for Min-max Optimization
Simons Institute: Learning & Games Reading Group , March 2022 • slides - Generative Adversarial Networks
Mediterranean Machine Learning (M2L) school, Jan. 2021 • video • slides • colab - An Introduction to JAX
Feb. 2020 • slides • colab - Reducing Noise in GAN Training with Variance Reduced Extragradient
Selected as oral presentation at Montreal AI Symposium and at ICML workshop on Generative Modeling and Model-Based Reasoning for Robotics and AI, 2019 • slides - Generative Adversarial Networks: Algorithm, Challenges & Research Topics
Paris GANs & co meetup, Oct. 2018 • slides - SGAN: An Alternative Training of Generative Adversarial Networks & Research Topics
Idiap/HES-SO AI workshop, April 2018 • slides
Teaching
- Games in Machine Learning (GML), winter 2023/24 @ Saarland University, advanced lecture.
The course covers:
- Game theory & algorithmic game theory basics
- Variational Inequalities (VIs) — extend classical convex optimization, enabling the formalization of a wide range of problems related to finding equilibria, such as robustified objectives, min-max and actor-critic problems, and multi-agent reinforcement learning, among others.
- Gradient-based optimization methods to solve VIs and their convergence.
- Some applications of VIs in machine learning, such as Generative Adversarial Networks.
See Course Materials
Games in ML Group
- Khaled Alomar, On solving Variational Inequalities with Gradient-based methods: Convergence Analysis internship since Nov. 2023, Math Dept. Saarland University.
- Baraah Adil Mohammed Sidahmed, On Solving Multi-Agent Reinforcement Learning with Optimization Methods for Equilibria MSc thesis & intern since Dec. 2023, Data Science and AI Dept., Saarland University.
- Sneha Chetani, On Robust Machine Learning internship since Feb. 2024, Data Science and AI Dept., Saarland University.
- Aniket Sanyal, On improving Variational Inequalities optimization methods using insights from Signal Processing internship since Feb. 2024, Computer Science Dept., Saarland University.
- Prashanth Pombala, On improving Variational Inequalities optimization methods using insights from continuous time internship Jan.-Apr. 2024, Mathematics and Computer Science Dept., Saarland University.
Supervision & Teaching Assistance (up to 2023)
- Tong Yang, On Interior Point Approach for Solving Variational Inequalities 2022, UC Berkeley.
- Gilberto Manunza (MSc thesis), On the Connection between Adversarial Training and Uncertainty Estimation seven months MSc project, 2021, EPFL.
- Apostolov Alexander (CS-498 Semester Project), On the Effect of Variance Reduced Gradient and Momentum for Optimizing Deep Neural Networks autumn semester, 2020, EPFL.
- Oğuz Kaan Yüksel, Normalizing Flows for Generalization and Robustness (CS-498 Semester Project) autumn semester, 2020, EPFL. Co-superivison with Sebastian U. Stich
- Yehao Liu, On the Drawbacks of Popular Deep Learning Uncertainty Estimation Methods spring semester & summer internship, 2021, EPFL. Co-superivison with Sebastian U. Stich
- Co-superivison with Mary-Anne Hartley:
- Deeksha M. Shama, Deep Learning Approaches for Covid-19 Diagnosis via Digital Lung Auscultation autumn semester, 2020.
- Pablo Cañas, On Uncertainty Estimation of Global COVID Policy Simulator autumn semester, 2020.
TAing
- Teaching Assistant, Deep Learning Course (EE-559) for MSc students, EPFL, 2018 & 2020.
- Teaching Assistant, An Introduction to Deep Learning for MSc students, African Master's in Machine Intelligence, Kigali, Rwanda, 2018.
Activities
Guest Editor
- Minimax Theory and its Applications. Special issue on min-max and Variational Inequality problems. Jointly with Gauthier Gidel, Manolis Zampetakis, Michael Jordan. Since 2024.
- Conferences: • ICLR • ICML • AISTATS • NeurIPS
- Journals: • JMLR • TMLR • SIAM Journal on Mathematics of Data Science • Springer Nature Machine Learning
- Workshops: • Optimization for Machine Learning (at NeurIPS)
- Advisory committee and WiML liaison for: WiML Un-Workshop @ ICML '21 and WiML Un-Workshop @ ICML '22. Held at the International Conference on Machine Learning (ICML). July, 2021 and July, 2022.
- Part of the Advisory Committee. ICCV Workshop on Multi-camera Multiple People Tracking, Oct., 2021.
- Co-organized WiML Virtual Un-Workshop as a Program Chair. Held at the International Conference on Machine Learning (ICML). July, 2020.
- Co-organized the WiML-CWS Social @ AISTATS 2022, on behalf of WiML. Held at the AISTATS. March, 2022.
- Co-organized the WiML-CWS Event: Community-Driven Mentoring Event and Panel, on behalf of WiML. Held at the AISTATS. April, 2021.
- Organizer of the Smooth Games Reading Group at EPFL. Oct. 2020 - March 2021.
- Vice President of Women in Machine Learning. Since April 2023.
- Board of Directors for Women in Machine Learning. Since April 2021.
- On Continuous time methods for ML workshop at ICML 2022. With Ricky Chen, Priya Donti, Adil Salim, and Michael Arbel. July 2022.
Reviewing
Workshops
Socials
Reading Group
Misc
Panel
Contact
If you'd like to discuss my work or broadly on doing research in CS, DEI activities, etc., please don't hesitate to reach out at: tatjana.chavdarova[at]berkeley[dot]edu