Activity stream from the Prorok Lab — papers, talks, videos, posts, and openings.https://prorok-lab.com/https://prorok-lab.com/publications/concrete-mapf-npj-robotics/https://prorok-lab.com/publications/concrete-mapf-npj-robotics/PAPER — Keisuke Okumura, Guang Yang, Zhan Gao, Heedo Woo, Amanda ProrokMon, 16 Mar 2026 00:00:00 GMThttps://prorok-lab.com/videos/concrete-mapf/https://prorok-lab.com/videos/concrete-mapf/Framework for aggressive multi-robot coordination deployed with 40 robots (20 aerial, 8 ground, 12 obstacles) in a compact lab space. Published in npj Robotics (2026).Mon, 16 Mar 2026 00:00:00 GMThttps://prorok-lab.com/publications/iclr2026-diversity-rewarded/https://prorok-lab.com/publications/iclr2026-diversity-rewarded/PAPER — Michael Amir, Matteo Bettini, Amanda ProrokTue, 20 Jan 2026 00:00:00 GMThttps://prorok-lab.com/publications/iclr2026-hypergraph-mapf/https://prorok-lab.com/publications/iclr2026-hypergraph-mapf/PAPER — Rishi Jain, Keisuke Okumura, Michael Amir, Pietro Lio, Amanda ProrokTue, 20 Jan 2026 00:00:00 GMThttps://prorok-lab.com/publications/iclr2026-watermarking/https://prorok-lab.com/publications/iclr2026-watermarking/PAPER — Michael Amir, Manon Flageat, Amanda ProrokTue, 20 Jan 2026 00:00:00 GMThttps://prorok-lab.com/publications/neurips2025-thompson-sampling/https://prorok-lab.com/publications/neurips2025-thompson-sampling/PAPER — Jasmine Bayrooti, Sattar Vakili, Amanda Prorok, Carl Henrik EkWed, 10 Dec 2025 00:00:00 GMThttps://prorok-lab.com/publications/extending-robot-minds/https://prorok-lab.com/publications/extending-robot-minds/PAPER — Amanda ProrokWed, 24 Sep 2025 00:00:00 GMThttps://prorok-lab.com/posts/science-robotics-extending-minds/https://prorok-lab.com/posts/science-robotics-extending-minds/It's finally here! My new Viewpoint Article 'Extending robot minds through collective learning' is published in Science Robotics. I argue for a paradigm shift: collective robotic intelligence — diverse, specialised agents that learn and work together — instead of massive monolithic 'do-it-all' generalist robots.Wed, 24 Sep 2025 00:00:00 GMThttps://prorok-lab.com/publications/recode-corl2025/https://prorok-lab.com/publications/recode-corl2025/PAPER — Michael Amir, Guang Yang, Zhan Gao, Keisuke Okumura, Heedo Woo, Amanda ProrokTue, 12 Aug 2025 00:00:00 GMThttps://prorok-lab.com/videos/recode-corl2025/https://prorok-lab.com/videos/recode-corl2025/Hybrid decentralized framework combining optimization-based control with MARL. Learns dynamic constraints to improve multi-agent coordination. Presented at CoRL 2025.Tue, 12 Aug 2025 00:00:00 GMThttps://prorok-lab.com/publications/d4orm-iros2025/https://prorok-lab.com/publications/d4orm-iros2025/PAPER — Yixiao Zhang, Keisuke Okumura, Heedo Woo, Ajay Shankar, Amanda ProrokFri, 04 Jul 2025 00:00:00 GMThttps://prorok-lab.com/videos/d4orm-iros2025/https://prorok-lab.com/videos/d4orm-iros2025/Optimization method generating kinodynamically feasible, collision-free multi-robot trajectories using diffusion denoising. Evaluated for teams of up to 16 robots in 2D and 3D worlds. Presented at IROS 2025.Fri, 04 Jul 2025 00:00:00 GMThttps://prorok-lab.com/publications/dvm-slam-icra2025/https://prorok-lab.com/publications/dvm-slam-icra2025/PAPER — Joshua Bird, Jan Blumenkamp, Amanda ProrokThu, 15 May 2025 00:00:00 GMThttps://prorok-lab.com/publications/language-conditioned-offline-rl-icra2025/https://prorok-lab.com/publications/language-conditioned-offline-rl-icra2025/PAPER — Steven Morad, Ajay Shankar, Jan Blumenkamp, Amanda ProrokThu, 15 May 2025 00:00:00 GMThttps://prorok-lab.com/publications/iclr2025-thompson-sampling/https://prorok-lab.com/publications/iclr2025-thompson-sampling/PAPER — Jasmine Bayrooti, Carl Henrik Ek, Amanda ProrokThu, 24 Apr 2025 00:00:00 GMThttps://prorok-lab.com/talks/ace-network-forum-oxford-2025/https://prorok-lab.com/talks/ace-network-forum-oxford-2025/Invited · ACE Network Forum — University of OxfordWed, 26 Mar 2025 00:00:00 GMThttps://prorok-lab.com/posts/ace-network-forum-oxford/https://prorok-lab.com/posts/ace-network-forum-oxford/Prorok Lab at the ACE Network Forum! Amanda Prorok delivered a keynote at the University of Oxford: 'Synthesizing Multi-Agent Policies: From Cooperative Robot Perception to Human-Led Fleet Control' — sharing our latest research on how multi-agent systems can enable smarter decision-making.Wed, 26 Mar 2025 00:00:00 GMThttps://prorok-lab.com/posts/bbc-click-feature/https://prorok-lab.com/posts/bbc-click-feature/Big news — our work has been featured on BBC Click! From swarm robotics to autonomous collaboration, our innovations are shaping the future of intelligent systems. Huge shoutout to Keisuke Okumura, Peter Woo, Guang Yang, Jan Blumenkamp, and Zhan Gao for their hard work.Mon, 17 Mar 2025 00:00:00 GMThttps://prorok-lab.com/publications/co-optimizing-environments-tro2025/https://prorok-lab.com/publications/co-optimizing-environments-tro2025/PAPER — Zhan Gao, Guang Yang, Amanda ProrokSat, 01 Mar 2025 00:00:00 GMThttps://prorok-lab.com/posts/hiring-ra-natural-language/https://prorok-lab.com/posts/hiring-ra-natural-language/We're hiring! Looking for a Research Assistant in Adaptive Multi-Robot Mission Control via Natural Language. Develop robot policies that respond to high-level natural language commands at the intersection of LLMs, NLP, and multi-robot systems.Tue, 04 Feb 2025 00:00:00 GMThttps://prorok-lab.com/openings/phd-perception-aerial-swarms/https://prorok-lab.com/openings/phd-perception-aerial-swarms/Develop robot policies that respond to high-level natural language commands at the intersection of LLMs, NLP, and multi-robot systems. Collaborate with students and researchers on cutting-edge projects in a world-class research environment.Tue, 04 Feb 2025 00:00:00 GMThttps://prorok-lab.com/publications/benchmarl-jmlr2024/https://prorok-lab.com/publications/benchmarl-jmlr2024/PAPER — Matteo Bettini, Amanda Prorok, Vincent MoensThu, 19 Dec 2024 00:00:00 GMThttps://prorok-lab.com/posts/neurips2024-benchmarl/https://prorok-lab.com/posts/neurips2024-benchmarl/Matteo Bettini alongside Vincent Moens from Meta's PyTorch team presented 'BenchMARL: Benchmarking Multi-Agent Reinforcement Learning' at NeurIPS 2024 — a cutting-edge library to simplify and standardize benchmarking for MARL, enabling seamless mixing of algorithms, tasks, and models.Thu, 19 Dec 2024 00:00:00 GMThttps://prorok-lab.com/talks/neurips2024-benchmarl/https://prorok-lab.com/talks/neurips2024-benchmarl/Conference · NeurIPS 2024Thu, 12 Dec 2024 00:00:00 GMThttps://prorok-lab.com/publications/provably-safe-navigation-corl2024/https://prorok-lab.com/publications/provably-safe-navigation-corl2024/PAPER — Zhan Gao, Guang Yang, Jasmine Bayrooti, Amanda ProrokWed, 06 Nov 2024 00:00:00 GMThttps://prorok-lab.com/publications/covis-net-corl2024/https://prorok-lab.com/publications/covis-net-corl2024/PAPER — Jan Blumenkamp, Steven Morad, Jennifer Gielis, Amanda ProrokFri, 18 Oct 2024 00:00:00 GMThttps://prorok-lab.com/videos/covis-net-corl2024/https://prorok-lab.com/videos/covis-net-corl2024/Decentralized visual spatial foundation model enabling real-time pose estimation and bird's-eye-view representations for multi-robot systems, without needing camera overlap. Presented at CoRL 2024.Fri, 18 Oct 2024 00:00:00 GMThttps://prorok-lab.com/posts/covis-net-corl2024/https://prorok-lab.com/posts/covis-net-corl2024/Excited to share our latest work: CoViS-Net, a decentralized visual spatial foundation model enabling real-time, platform-agnostic pose estimation and spatial comprehension for autonomous robots. Unlike classical methods, CoViS-Net provides accurate pose estimates and a local BEV without needing camera overlap.Wed, 16 Oct 2024 00:00:00 GMThttps://prorok-lab.com/talks/eth-zurich-rss-seminar/https://prorok-lab.com/talks/eth-zurich-rss-seminar/Invited · ETH Zürich — Distinguished Seminar in Robotics, Systems and ControlFri, 04 Oct 2024 00:00:00 GMThttps://prorok-lab.com/posts/eth-zurich-distinguished-seminar/https://prorok-lab.com/posts/eth-zurich-distinguished-seminar/Honored to deliver a lecture at the Distinguished Seminar in Robotics, Systems and Control at ETH Zurich — presenting on 'Synthesizing Multi-Agent Policies with Graph Neural Networks: From Cooperative Robot Perception to Human-Led Fleet Control'.Thu, 03 Oct 2024 00:00:00 GMThttps://prorok-lab.com/publications/behavioral-diversity-icml2024/https://prorok-lab.com/publications/behavioral-diversity-icml2024/PAPER — Matteo Bettini, Ryan Kortvelesy, Amanda ProrokSun, 21 Jul 2024 00:00:00 GMThttps://prorok-lab.com/talks/tedx-oxford-2024/https://prorok-lab.com/talks/tedx-oxford-2024/Invited · TEDxOxfordMon, 06 May 2024 00:00:00 GMThttps://prorok-lab.com/publications/cambridge-robomaster-dars2024/https://prorok-lab.com/publications/cambridge-robomaster-dars2024/PAPER — Jan Blumenkamp, Steven Morad, Jennifer Gielis, Amanda ProrokWed, 01 May 2024 00:00:00 GMThttps://prorok-lab.com/posts/steven-morad-deep-rl-course/https://prorok-lab.com/posts/steven-morad-deep-rl-course/It's an absolute delight seeing PhD students progress to senior teaching roles. Steven Morad delivered an excellent Deep Reinforcement Learning course to MPhil and Part III students at Cambridge — can't imagine a better way to introduce Deep RL and its practical intricacies.Mon, 26 Feb 2024 00:00:00 GMThttps://prorok-lab.com/videos/so2-downwash/https://prorok-lab.com/videos/so2-downwash/Learning-based approach modelling downwash forces for multirotor formation flight, improving trajectory tracking by 36%. Published in IEEE RA-L 2023.Thu, 21 Dec 2023 00:00:00 GMThttps://prorok-lab.com/publications/f-mean-aggregation-neurips2023/https://prorok-lab.com/publications/f-mean-aggregation-neurips2023/PAPER — Ryan Kortvelesy, Steven Morad, Amanda ProrokSun, 10 Dec 2023 00:00:00 GMThttps://prorok-lab.com/publications/fast-forgetful-memory-neurips2023/https://prorok-lab.com/publications/fast-forgetful-memory-neurips2023/PAPER — Steven Morad, Ryan Kortvelesy, Stephan Liwicki, Amanda ProrokSun, 10 Dec 2023 00:00:00 GMThttps://prorok-lab.com/videos/aggregate-downwash/https://prorok-lab.com/videos/aggregate-downwash/Experimental study of propeller downwash interaction patterns in close-proximity multirotor formations. Presented at ISER 2023.Tue, 07 Nov 2023 00:00:00 GMThttps://prorok-lab.com/videos/docking-multirotors/https://prorok-lab.com/videos/docking-multirotors/Demonstrates successful physical docking between two airborne multirotors using learned downwash compensation. Presented at ISER 2023.Tue, 07 Nov 2023 00:00:00 GMThttps://prorok-lab.com/posts/neurips2023-fast-forgetful-memory/https://prorok-lab.com/posts/neurips2023-fast-forgetful-memory/Check out our new NeurIPS 2023 paper — 'Reinforcement Learning with Fast and Forgetful Memory'. A super-efficient plug-in replacement for RNNs/transformers that runs up to two orders of magnitude faster with linear space complexity. Sets new records on the POPGym benchmark.Mon, 16 Oct 2023 00:00:00 GMThttps://prorok-lab.com/videos/see-what-robot-cant-see/https://prorok-lab.com/videos/see-what-robot-cant-see/Graph Neural Network approach achieving up to 2.0x improvement in SPL through sensor-robot communication for cooperative visual navigation. Presented at IROS 2023.Mon, 31 Jul 2023 00:00:00 GMThttps://prorok-lab.com/publications/heterogeneous-multi-robot-rl/https://prorok-lab.com/publications/heterogeneous-multi-robot-rl/PAPER — Matteo Bettini, Ajay Shankar, Amanda ProrokFri, 05 May 2023 00:00:00 GMThttps://prorok-lab.com/videos/heterogeneous-marl/https://prorok-lab.com/videos/heterogeneous-marl/HetGPPO framework enabling heterogeneous behaviours in multi-agent reinforcement learning with graph neural networks. Presented at AAMAS 2023.Fri, 05 May 2023 00:00:00 GMThttps://prorok-lab.com/publications/popgym-iclr2023/https://prorok-lab.com/publications/popgym-iclr2023/PAPER — Steven Morad, Ryan Kortvelesy, Matteo Bettini, Stephan Liwicki, Amanda ProrokMon, 01 May 2023 00:00:00 GMThttps://prorok-lab.com/videos/vmas-simulator/https://prorok-lab.com/videos/vmas-simulator/PyTorch-based vectorized simulator achieving over 100x faster execution than comparable multi-agent frameworks. Presented at DARS 2022.Thu, 07 Jul 2022 00:00:00 GMThttps://prorok-lab.com/videos/explanation-aware-replay/https://prorok-lab.com/videos/explanation-aware-replay/Reinforcement learning approach leveraging explanations for improved experience replay in environments with many rules. Published in IEEE RA-L 2022.Mon, 27 Jun 2022 00:00:00 GMThttps://prorok-lab.com/talks/eth-zurich-autonomy-talks-2021/https://prorok-lab.com/talks/eth-zurich-autonomy-talks-2021/Invited · ETH Zürich — Autonomy TalksMon, 06 Dec 2021 00:00:00 GMThttps://prorok-lab.com/videos/gnn-real-world-framework/https://prorok-lab.com/videos/gnn-real-world-framework/ROS2-based framework demonstrating the first real-world deployment of graph neural network policies on decentralized multi-robot systems. Presented at ICRA 2022.Wed, 03 Nov 2021 00:00:00 GMThttps://prorok-lab.com/videos/modgnn/https://prorok-lab.com/videos/modgnn/Generalised GNN framework improving performance and multi-agent task generalisation. Presented at ICRA 2021.Fri, 26 Mar 2021 00:00:00 GMT