Workshop Recording Available: https://www.youtube.com/watch?v=33JbcVqnVEM <https://www.youtube.com/watch?v=33JbcVqnVEM&t=7463s> &t=7463s
Dear roboticists, our workshop Machine Learning for Motion Planning (https://sites.google.com/utexas.edu/mlmp-icra2021) successfully took place in a virtual format on May 31 2021. The recording of the entire workshop is available online: https://www.youtube.com/watch?v=33JbcVqnVEM <https://www.youtube.com/watch?v=33JbcVqnVEM&t=7463s> &t=7463s Please feel free to distribute the recording with the community you think may be interested. Thanks Xuesu -- Xuesu Xiao, Ph.D. Postdoctoral Researcher Department of Computer Science The University of Texas at Austin GDC 3.418 +1 (512) 471-9765 x...@cs.utexas.edu <mailto:x...@cs.utexas.edu> https://www.cs.utexas.edu/~xiao/ Machine Learning for Motion Planning Call for Participation Workshop Location: https://utexas.zoom.us/j/91637134883 Workshop Date: May 31 2021 Workshop Website: <https://sites.google.com/utexas.edu/mlmp-icra2021> https://sites.google.com/utexas.edu/mlmp-icra2021 Submission Site: <https://easychair.org/conferences/?conf=mlmp2021> https://easychair.org/conferences/?conf=mlmp2021 Submission Deadline: May 7th 2021 Motion planning is one of the core problems in robotics with applications ranging from navigation to manipulation in complex cluttered environments. It has a long history of research with methods promising full to probabilistic completeness and optimality guarantees. However, challenges still exist when classical motion planners face real-world robotics problems in high dimensional or highly constrained workspaces. The community continues to develop new strategies to overcome limitations associated with these methods, which include computational and memory burdens, planning representation, and the curse of dimensionality. In contrast, recent advancements in machine learning have opened up new perspectives for roboticists to look at the motion planning problem: bottlenecks of classical motion planners can be addressed in a data-driven manner; classical planners can go beyond the geometric sense and enable orthogonal planning capabilities, such as planning with visual or semantic input, or in a socially-compliant manner. The objective of this workshop is to bring the two research communities under one forum to discuss the lessons learned, open questions, and future directions of machine learning for motion planning. We aim to identify the gaps and formalize the merging points between the two schools of methodologies, e.g. workspace representation, sample generation, collision checking, cost definition, and answer the questions of why, where, and how to apply machine learning for motion planning. Papers of up to two-six pages are sought in the following topic areas: Topics of interest: * Data-driven approaches to motion planning * Learning-based adaptive sampling methods * Learning models for planning and control * Imitation learning for planning and control * Learning generalizable and transferable planning models * Representation learning for planning * Learning-based collision detection, edge selection, and pruning techniques, and related topics * Data-efficiency in data-driven techniques to planning * Formal guarantees to machine learning-based planning methods * Learning methods for hierarchical planning such task and motion planning, multi-model motion planning, and related topics * Active/lifelong/continual learning methods for planning and related topics Organizers: 1. Xuesu Xiao, Department of Computer Science, The University of Texas at Austin, 2317 Speedway, Austin, TX 78712, USA, Phone: +1 (512) 471-9765, Email: <mailto:x...@cs.utexas.edu> x...@cs.utexas.edu, URL: <https://www.cs.utexas.edu/~xiao/> https://www.cs.utexas.edu/~xiao/ (Primary Contact) 2. Ahmed H. Qureshi, Department of Electrical and Computer Engineering, University of California San Diego, 9500 Gilman Dr, La Jolla, CA 92093, USA, Phone: +1 (858) 349-8122, Email: <mailto:a1qur...@ucsd.edu> a1qur...@ucsd.edu, URL: <https://qureshiahmed.github.io/> https://qureshiahmed.github.io/ 3. Anastasiia Varava, School of Computer Science and Communication, KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden, Email: <mailto:var...@kth.se> var...@kth.se, URL: <https://anvarava.github.io/> https://anvarava.github.io/ 4. Michael Everett, Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, 77 Massachusetts Ave, 31-235C, Cambridge, MA 02139, Phone: +1 (734) 476-2051, Email: <mailto:m...@mit.edu> m...@mit.edu, URL: <http://mfe.mit.edu> http://mfe.mit.edu 5. Michael C. Yip, Department of Electrical and Computer Engineering, University of California San Diego, 9500 Gilman Dr, La Jolla, CA 92093, USA, Phone: +1 (858) 822-4778, Email: <mailto:y...@ucsd.edu> y...@ucsd.edu, URL: <https://yip.eng.ucsd.edu/> https://yip.eng.ucsd.edu/ 6. Peter Stone, Department of Computer Science, The University of Texas at Austin, 2317 Speedway, Austin, TX 78712, USA, Phone: +1 (512) 471-9796, Email: <mailto:pst...@cs.utexas.edu> pst...@cs.utexas.edu, URL: <https://www.cs.utexas.edu/~pstone/> https://www.cs.utexas.edu/~pstone/ Steering Committee: 1. Danica Kragic, KTH Royal Institute of Technology, Sweden. Email: <mailto:d...@kth.se> d...@kth.se 2. Jonathan How, Massachusetts Institute of Technology (MIT), USA. Email: 3. Jan Peters, Technische Universität Darmstadt, Germany. Email: <mailto:pet...@ias.tu-darmstadt.de> pet...@tu-darmstadt.de 4. Howie Choset, Carnegie Mellon University (CMU), USA. Email: <mailto:cho...@cmu.edu> cho...@cmu.edu 5. Steven LaValle, University of Oulu, Finland. Email: <mailto:steven.lava...@oulu.fi> steven.lava...@oulu.fi 6. Lydia Kavraki, Rice University, USA. Email: <mailto:kavr...@rice.edu> kavr...@rice.edu 7. Seth Hutchinson, GeorgiaTech, USA. Email: <mailto:s...@gatech.edu> s...@gatech.edu 8. Aude Billard, École polytechnique fédérale de Lausanne (EPFL), <mailto:aude.bill...@epfl.ch> aude.bill...@epfl.ch 9. Aleksandra Faust, Google Brain Research, <mailto:fa...@google.com> fa...@google.com Invited Speakers: 1. Sertac Karaman, Massachusetts Institute of Technology (MIT). Email: <mailto:ser...@mit.edu> ser...@mit.edu 2. Raquel Urtasun, University of Toronto & Uber ATG. Email: <mailto:urta...@cs.toronto.edu> urta...@cs.toronto.edu 3. Marc Toussaint, Technische Universität Berlin. Email: <mailto:toussa...@tu-berlin.de> toussa...@tu-berlin.de 4. Anca Dragan, University of California Berkeley, USA. Email: <mailto:a...@berkeley.edu> a...@berkeley.edu Schedule (GMT-04): 09:00-09:05 Opening Remarks 09:05-09:35 Invited Talk: Marc Toussaint 09:35-09:50 Neural Network Based Model Predictive Control for an Autonomous Vehicle 09:50-10:05 Deep Neural Network-based Fast Motion Planning Framework for Quadrupedal Robot 10:05-10:25 SPOTLIGHT: Unsupervised Path Regression Networks 10:25-10:30 Coffee Break 10:30-10:45 Continuous Robot Navigation in Unknown Environments Using Deep Reinforcement Learning 10:45-11:15 Invited Talk: Raquel Urtasun 11:15-11:30 DMP Based Perturbed Handover with Preferential Learning 11:30-12:30 Lunch Break 12:30-01:00 Invited Talk: Anca Dragan 01:00-01:15 High-Speed Drone Flight with On-Board Sensing and Computing 01:15-01:30 Learning Near-Time-Optimal Quadrotor Trajectories 01:30-01:45 Improving End-to-End Drone Racing by Visual Attention Prediction 01:45-01:55 Coffee Break 01:55-02:15 SPOTLIGHT: PathBench: A Benchmarking Platform for Classical and Learned Path Planning Algorithms 02:15-02:30 Waypoint Planning Networks 02:30-02:50 SPOTLIGHT: Motion Planning via Bayesian Learning in the Dark 02:50-03:00 Coffee Break 03:00-03:15 Motion Planner Guided Visuomotor Policy Learning 03:15-03:30 PRX: a Light and Flexible Library for Development and Benchmarking of Sample-Based Kinodynamic Planners 03:30-03:45 Data-Efficient Learning of High-Quality Controls for Kinodynamic Planning used in Vehicular Navigation 03:45-04:00 Coffee Break 04:00-04:30 Invited Talk: Sertac Karaman 04:30-05:00 Discussion & Closing RemarksTechnical Committee Endorsement: 1. IEEE RAS Technical Committee on Algorithms for Planning and Control of Robot Motion 2. IEEE-RAS Technical Committee on Robot Learning For questions, please contact Dr. Xuesu Xiao Department of Computer Science The University of Texas at Austin 2317 Speedway, Austin, Texas 78712-1757 USA +1 (512) 471-9765 x...@cs.utexas.edu <mailto:x...@cs.utexas.edu> https://www.cs.utexas.edu/~xiao/ -- Xuesu Xiao, Ph.D. Postdoctoral Researcher Department of Computer Science The University of Texas at Austin GDC 3.418 +1 (512) 471-9765 x...@cs.utexas.edu <mailto:x...@cs.utexas.edu> https://www.cs.utexas.edu/~xiao/
_______________________________________________ uai mailing list uai@engr.orst.edu https://it.engineering.oregonstate.edu/mailman/listinfo/uai
