Thinh T. Doan

An educator and researcher working as an Assistant Professor at @UTAerospace.

[Thinh T. Doan @thinh_t_doan]

I am humbled and delighted to receive this award. I am deeply indebted to the tremendous and constant support of my advisors and mentors, Carolyn Beck and R Srikant, as well as my students, colleagues, friends, and my department.

https://t.co/JX5VUBL0hx

[Thinh T. Doan @thinh_t_doan]

We are excited to explore TAG as a platform for connecting learning algorithms with real-world physical dynamics and investigating new challenges in learning, control, and optimization for multi-agent systems.

Feedback and comments are very welcome!

https://t.co/wGK6Yo2MuQ

[Thinh T. Doan @thinh_t_doan]

Excited to introduce our new testbed: TAG — Testbed for Autonomous Games.
TAG is a reproducible physical platform for developing and testing AI and control algorithms that must learn, predict, and make decisions in the presence of physical coupling and dynamic environments.

[Thinh T. Doan @thinh_t_doan]

@bremen79 I had one paper rejected last year. Glad to know that it's due to the rules, what ever they are. Feeling much better now. Lol.

[Thinh T. Doan @thinh_t_doan]

The match will be tentatively held on Wednesday December 10th (the second to last day of CDC) in the afternoon/evening. We will likely to have a field during the evening time to avoid the conflict with the talks and the heat in Rio. It is an outdoor field.

[Thinh T. Doan @thinh_t_doan]

Friends and colleagues:

We're excited to announce the 2025 CDC Cup for faculty, where we will have a friendly match of soccer/football in Rio. It’s Brazil so it will be a lot of fun.
Please register here so we can order the jerseys: https://t.co/VCRxTB1uxt

@IEEECDC2025

[Thinh T. Doan @thinh_t_doan]

IEEE CDC 2024 soccer dream teams. @IEEECDC2024

[Texas ASE/EM @UTAerospace]

The sky is not the limit with our @UTAerospace graduate programs. Deadline to apply is Dec. 1! 🚀🤘

[Thinh T. Doan @thinh_t_doan]

@MUBNiazi Congrats Umar. Very happy for you. Let's celebrate next month in Milan.

[Thinh T. Doan @thinh_t_doan]

Prospective graduate students: Come to join our department! It's a great place for you to spend in the next few years. My lab is also looking for new students who are passionate about control, optimization, and online learning for autonomy. https://t.co/TPJLwYRJO9

[Texas ASE/EM @UTAerospace]

🚀Ready reach new heights? The deadline to apply for our top-ranked graduate programs is Dec. 1! Our students work on research related to aviation, space engineering and science, robotics, energy, earth science, mechanics of materials, biomedicine + more: https://t.co/bMCAJmQEMj

[Thinh T. Doan @thinh_t_doan]

@SGrm87 It sounds like a great conference to attend. Wish I'd have known it.

[ACC2026 @ACC_Control]

🚨 Reminder: The deadline for submitting nominations for 2025 ACC Student Best Paper Award is fast approaching! ⏳ Make sure to submit your nomination form by October 14th, 2024 to recognize outstanding student research. Don’t miss the chance to highlight innovative work! 🌟.

[Thinh T. Doan @thinh_t_doan]

We call the year of AI.

https://t.co/AtvCChARxQ

[Texas ASE/EM @UTAerospace]

We have 2 new tenure-track faculty positions open @UTAerospace!

🌟Assistant Professor in Emerging Aeronautics
🌟Assistant Professor in Space Systems

🚀 Help us shape the unlimited future of space in one of the best U.S. cities to live! https://t.co/2ih2BtICF1

[Thinh T. Doan @thinh_t_doan]

@beenwrekt Very nice interpretation.
In a (ir)relevant source, @GeorgeLan7 also uses game theory (or duality) to beautifully explain Nesterov accelerated gradient (Section 3.4 of his book). He uses this idea quite a bit in his research.

[Thinh T. Doan @thinh_t_doan]

Join our workshop on safe and secure learning systems to learn the cutting edge research from leading experts in this area at CDC 2024 @IEEECDC2024. Hope to see every one in Milan.

[Thinh T. Doan @thinh_t_doan]

Together with Kyriakos Vamvoudakis from GaTech, we are organizing a Workshop on Safe and Secure Learning-Enabled Systems at 2024 Conference on Decision and Control, Milan, Italy. @IEEECDC2024
https://t.co/Ur99bZxZBy

[Thinh T. Doan @thinh_t_doan]

Our newest work on the optimal complexity of multi-time-scale stochastic approximation is up. With this new algorithm, we can scale up to any number of equations without decaying the convergence rates. Any comments would be highly appreciated.

[arXiv math.OC Optimization and Control @mathOCb]

Sihan Zeng, Thinh T. Doan: Accelerated Multi-Time-Scale Stochastic Approximation: Optimal Complexity and Applications in Reinforcement Learning and Multi-Agent Games https://t.co/o7HGSMrDvk https://t.co/yN2OQwf06j

[Thinh T. Doan @thinh_t_doan]

Key observation: When the local functions are sufficiently redundant but can be heterogeneous, one can show that local SGD is resilient to Byzantine malicious and achieves an optimal convergence rate to the optimal solution.

[Thinh T. Doan @thinh_t_doan]

Our newest paper, leaded by my PhD student Amit Dutta, on exact fault-tolerance in Byzantine federated learning is up. Multi-time-scale SA is again the key idea in stochastic settings. Any comments would be highly appreciated.

[arXiv math.OC Optimization and Control @mathOCb]

Amit Dutta, Thinh T. Doan: Resilient Two-Time-Scale Local Stochastic Gradient Descent for Byzantine Federated Learning https://t.co/cqaxcuwMNV https://t.co/35xxYASye6

[Thinh T. Doan @thinh_t_doan]

@UCIEngineerDean @UCIEngineering @oai_uci Great run-up for CDC cup in Milan :)