Invitation to Multi-Agent Reinforcement Learning (AI Agents) Reading Group at UMD, College Park

[Saptarashmi Bandyopadhyay]

Hey UMD CS and ECE students and everyone,

Hope all of you are having a great Summer and are looking forward to the Fall 2024 Semester.

Happy to share about our Reading Group and Seminar Series on Multi-Agent Reinforcement Learning (MARL) and Autonomous Decision Making, for real-world AI Agents. 

The Reading Group is organized by me and co-organized by Ryan Sullivan. We are 4th year CS PhD Candidates advised by Dr. John Dickerson https://www.cs.umd.edu/people/dickerson. I am a Student Researcher Intern at Google AI AR (previously an AI Resident at GoogleX) working on AI Agents and Ryan is a Research Intern at SonyAI (previously a Student Researcher at Google Research) working on Curriculum Learning. 

Our Team includes Jack Cole, rising MS student in Computer Science (BS CS alum), currently a Software Engineer, OriginAI and Associate Team Member, Anirudh Satheesh, BS student in Computer Science and Sony Intern.

If you're interested in the domain of Artificial Intelligence (AI), Autonomous Decision Making exploring Multi-Agent Learning, Multimodal Learning, Meta-Learning, Reinforcement Learning, Imitation Learning, Self-Supervised Learning, Compositional Learning, Curriculum Learning, Predictive Control, In-context Learning, Game Theoretic insights in AI and other AI paradigms, please fill up the google form on our website for more https://sites.google.com/view/university-of-maryland-marl/

Our website can also be found at https://go.umd.edu/marl. We have 900+ participants from 6 continents.

Previous and upcoming speakers (as of now) include Dr. John Langford from Microsoft Research, NYC; Dr. Aleksandra Faust from Google Brain, Dr. Marc Lanctot, Dr. Julien Perolat, Dr. Edward Hughes, Dr. Luke Marris, Dr. Roma Patel, Dr. Vikranth Dwaracherla, Dr. Ruofei Du, Dr. Prateek Jain, Dr. Petar Velickovic  from Google DeepMind, Dr. Stephen McAleer and Brandon Trabucco from Carnegie Mellon University, Dr. Kai Wang from Harvard University (now at Georgia Tech), Dr. Chi Jin from Princeton University, Dr. Robert Loftin from TU Delft (formerly at Microsoft Research), Dr. Branislav Kveton from Amazon AWS AI Labs, Dr. Thodoris Lykouris from MIT, Dr. Eugene Vinitsky from Apple/NYU/UC Berkeley, Joseph Suarez and Chanwoo Park from Massachusetts Institute of Technology (MIT) EECS,Chris Lu from Oxford University, Stephanie Milani, Dr. Andrea Bajcy, Dr. Ben Eysenbach from Carnegie Mellon University (now at Princeton University), Samuel Sokota from Carnegie Mellon University,  Caroline Wang from University of Texas, Austin, Tony Zhao and Andy Shih from Stanford University, Dr. Jordan Terry CEO of Farama Foundation, Mikayel Samvelyan from Meta AI & University of College London (UCL) and UMD CS PhD alumnus and Niklas Lauffer from UC Berkeley.

The aim is to pursue foundational research discussions, brainstorming and potential collaboration on MARL and autonomous decision making frameworks with application areas like self-driving cars, solving game theory problems like discovering, analyzing, disrupting and interdicting illicit trafficking networks (be it arms, humans, wildlife, rare plants and deforested trees, illegal drugs, counterfeit goods), war games and strategies like the Diplomacy Game, equitable markets, fair tax policies, climate change and existential disasters like pandemics, climate conservation (deforestation mitigation, anti-poaching), global economic orchestration like supply chains, autonomous asset management like stock portfolios, multi-modal & multi-lingual LLMs(Large Language Models)/VLMs(Vision Language Models)/DT(Decision transformers)/Robotics Transformers(RT)/Multimodal Transformers, Explainable AI (XAI), Interpretable AI, autonomous scheduling, smart power grids, swarm robotics, agriculture planning among other domains. We consider the fundamental Human-AI alignment problem through Multi-Agent RL to be pertinent for discussions in our Reading Group.

We are building an Open Source Multi-Agent AI Research Community with students and researchers experienced in AI Agents research. The goals of our open-source Multi-Agent AI Community are to promote:

1. Open-source code and Multi-Agent AI models (Like publishing on Github)

2. Research publications at reputed conferences (AAAI, KDD, ICRA, AAMAS, ICLR, CVPR, ACL, ICML etc.) and good journals (JMLR, TMLR) which allow open-source research

3. Compute efficient solutions to Multi-Agent AI problems in the open source to reduce compute dependency of Machine Learning models

4. Networking in a Dynamic Multi-Agent AI community

Upcoming talks include Petar Velickovic from Google DeepMind on July 30, Ruofei Du from Google on August 6 and Prateek Jain from Google DeepMind on September 3. More talks will be added soon. Details are on our website https://sites.google.com/view/university-of-maryland-marl/home?authuser=0

Usually, we organize the Reading group every Tuesday at 6:30 pm EST (after class hours) virtually on Google Meet and in-person at IRB-5105.

Timings often change based on the availability of speakers in different time zones and are accordingly updated on our website https://go.umd.edu/marl

Please let me know if you have any questions.

Best Regards,

Saptarashmi Bandyopadhyay

4th Year PhD Candidate of Computer Science

University of Maryland, College Park

Student Researcher Intern at Google AI AR