Christopher Amato

Articles

Christopher Amato is an American computer scientist and Associate Professor in the Khoury College of Computer Sciences at Northeastern University in Boston, Massachusetts. His research focuses on reinforcement learning, multi-Agent systems, and planning in partially observable environments. He leads the Lab for Learning and Planning in Robotics (LLPR).

Education

Amato received a Bachelor of Arts in Clinical Psychology and Philosophy from Tufts University. He pursued graduate studies at the University of Massachusetts Amherst, earning a Master of Science in Computer Science followed by a Ph.D. in Computer Science under the supervision of Shlomo Zilberstein. His dissertation focused on increasing scalability in algorithms for centralized and decentralized partially observable Markov decision processes (POMDPs).

Career

After completing his doctorate, Amato worked as a research scientist at Aptima, Inc. He subsequently held positions as a postdoctoral fellow and research scientist at the Massachusetts Institute of Technology (MIT), where he worked with Leslie P. Kaelbling and the Learning and Intelligent Systems group in the Computer Science and Artificial Intelligence Laboratory (CSAIL), as well as Jonathan P. How and the Aerospace Control Lab in the Laboratory for Information and Decision Systems (LIDS). He also served as an Assistant Professor at the University of New Hampshire from 2015 to 2016.

Amato joined Northeastern University in 2016 as an Assistant Professor and was promoted to Associate Professor in 2022. At Northeastern, he leads the Lab for Learning and Planning in Robotics, focusing on reinforcement learning and planning in partially observable and multi-agent/multi-robot systems.

Research

Amato's research lies at the intersection of artificial intelligence, machine learning, and robotics. His work explores principled solution methods for systems of agents (such as robots, network nodes, sensors) operating under uncertainty with limited communication. Applications of his research include multi-robot navigation, search and rescue operations, autonomous surveillance, and multi-robot coordination.

His main research areas include:

  • Reinforcement learning (partially observable and multi-agent)
  • Multi-agent and multi-robot systems
  • Partially observable Markov decision processes (POMDPs)
  • Decentralized POMDPs (Dec-POMDPs)
  • Planning under uncertainty
  • Safe reinforcement learning
  • Adversarial machine learning in reinforcement learning contexts

Amato has published extensively in leading artificial intelligence, machine learning, and robotics venues including NeurIPS, ICML, ICLR, AAAI, IJCAI, AAMAS, ICRA, and IROS. According to Google Scholar, his work has been cited over 7,900 times.

Notable contributions

Amato has made significant contributions to the theory and application of decentralized decision-making under uncertainty:

  • Decentralized POMDPs: Amato is a leading researcher in the area of decentralized partially observable Markov decision processes (Dec-POMDPs), a framework for modeling multi-agent decision-making under uncertainty. He maintains the Dec-POMDP resource page, which serves as a central repository for publications, code, and datasets in this area.
  • Multi-agent reinforcement learning: His work on deep decentralized multi-task multi-agent reinforcement learning under partial observability (ICML 2017) has been highly influential, with over 590 citations, addressing the challenge of learning in multi-agent systems with limited communication.
  • Centralized training for decentralized execution (CTDE): Amato has contributed foundational work on the CTDE paradigm for cooperative multi-agent reinforcement learning.
  • Safe reinforcement learning: Recent work includes developing methods for shield decomposition for safe reinforcement learning in partially observable multi-agent environments.
  • Adversarial robustness in RL: His research group has investigated vulnerabilities of deep reinforcement learning to backdoor poisoning attacks, publishing notable works including "SleeperNets" (NeurIPS 2024) and "Adversarial Inception Backdoor Attacks" (ICML 2025).

Books

Amato has co-authored two books on multi-agent decision-making:

Awards and honors

  • Stanford University Top 2% Most-Cited Scientists – Included in the annual assessment of author citations representing the top 2% of most-cited scientists worldwide based on single-year impact.
  • NSF CAREER Award (2021) – Awarded by the National Science Foundation for research on "Scalable Planning and Learning for Multi-Agent Systems under Uncertainty."
  • Amazon Research Awards (2019, 2020) – For research on reinforcement learning and multi-agent systems.
  • Best Paper Award, AAMAS (2014) – International Conference on Autonomous Agents and Multiagent Systems.
  • Member, AAAI – Association for the Advancement of Artificial Intelligence.
  • Member, ACM – Association for Computing Machinery.