Knowledge, Skill, and Behavior Transfer in Autonomous Robots - 2014

Update

The schedule is now available.

Overview

Autonomous robots have achieved high levels of performance and reliability at specific tasks. However, for them to be practical and effective at everyday tasks in our homes and offices, they must be able to learn to perform different tasks over time, and rapidly adapt to new situations.

Learning each task in isolation is an expensive process, requiring large amounts of both time and data. In robotics, this expensive learning process also has secondary costs, such as energy usage and joint fatigue. Furthermore, as robotic hardware evolves or new robots are acquired, these robots must be trained, which is extremely inefficient if performed tabula rasa.

Recent developments in knowledge representation, machine learning, and optimal control provide a potential solution to this problem, enabling robots to minimize the time and cost of learning new tasks by building upon knowledge acquired from other tasks or by other robots. This ability is essential to the development of versatile autonomous robots that can perform a wide variety of tasks and rapidly learn new abilities.

Various aspects of this problem have been addressed by different communities in artificial intelligence and robotics. This symposium will seek to draw together researchers from these different communities toward the goal of enabling autonomous robots to support a wide variety of tasks, rapidly and robustly learn new abilities, adapt quickly to changing contexts, and collaborate effectively with other robots and humans.

Topics

The symposium will include paper presentations, talks, and discussions on a variety of topics related to lifelong learning, including but not limited to:

• Transfer in Autonomous Robots
• • Inter-Task Transfer Learning
  • Transfer Over Long Sequences of Tasks
  • Cross-Domain Transfer Learning
  • Long-Term Autonomy
  • Autonomy in Dynamic and Noisy Environments
  • Lifelong Learning
  • Knowledge Representation
  • Simulated to Real Robot Transfer, and Vice Versa
    
• Multi-Robot Systems
• • Multi-Robot Knowledge Transfer
  • Task Switching in Multi-Robot Learning
  • Distributed Transfer Learning
  • Knowledge/Skill Transfer Across Heterogeneous Robots
• Human-Robot Interaction
• • Human-Robot Knowledge/Skill Transfer
  • Knowledge/Skill Transfer in Mixed Human-Robot Teams
  • Learning by Demonstration, Imitation Learning
• Cloud Networked Robotics
• •  Access to Shared Knowledge, Reasoning, and Skills in the Cloud 
  •  Cloud-based Knowledge/Skill Transfer 
  • Cloud-based Distributed Transfer Learning
• Applications
• • Testbeds and Environments
  • Data Sets
  • Evaluation Methodology
    

Invited Speakers

• Yiannis Demiris, Imperial College London.
• Maja Mataric, University of Southern California.
• Stefan Schaal, University of Southern California.
• Peter Stone, University of Texas at Austin.
• Andrea Thomaz, Georgia Tech.
• Manuela Veloso, Carnegie Mellon University.

List of accepted papers

• Gabriel Barth-Maron, David Abel, James MacGlashan and Stefanie Tellex. Affordances as Transferable Knowledge for Planning Agents.
• Marie desJardins, Tenji Tembo, Nicholay Topin, Michael Bishoff, Shawn Squire, James MacGlashan, Rose Carignan and Nicholas Haltmeyer. Discovering Subgoals in Complex Domains.
• Stéphane Doncieux. Knowledge Extraction from Learning Traces in Continuous Domains.
• Gabriel Ferrer. Towards Human-Induced Vision-Guided Robot Behavior.
• Tesca Fitzgerald, Ashok Goel and Andrea Thomaz. Representing Skill Demonstrations for Adaptation and Transfer.
• Sachithra Hemachandra, Mattew Walter and Seth Teller. Information Theoretic Question Asking to Improve Spatial Semantic Representations.
• Bruce Johnson, Hairong Qi and Jason Isaacs. Computing a Heuristic Solution to the Watchman Route Problem by Means of Photon Mapping Within a 3D Virtual Environment Testbed.
• George Konidaris and Finale Doshi-Velez. Hidden Parameter Markov Decision Processes: An Emerging Paradigm for Modeling Families of Related Tasks.
• Milad S. Malekzadeh, Sylvain Calinon, Danilo Bruno and Darwin G. Caldwell. A Skill Transfer Approach for Continuum Robots - Imitation of Octopus Reaching Motion with the STIFF-FLOP Robot.
• Benjamin Rosman. Behavioural Domain Knowledge Transfer for Autonomous Agents.

Schedule

Please refer to the schedule page.

Submissions

Submissions are now closed.

Organizers

Matteo Leonetti (chair), University of Texas at Austin.
Eric Eaton (co-chair), University of Pennsylvania.
Pooyan Fazli (co-chair), Carnegie Mellon University.

Program Committee

Haitham Bou Ammar, University of Pennsylvania; Sylvain Calinon, Idiap Research Institute; Bruno Castro Da Silva, University of Massachusetts Amherst; Chris Clingerman, University of Pennsylvania; George Konidaris, Duke University; Daniele Magazzeni, King's College London; Dave Meger, McGill University; Tekin Mericli, Carnegie Mellon University; Cetin Mericli, Carnegie Mellon University; David Portugal, University of Coimbra; Paul Ruvolo, Olin College; Jivko Sinapov, University of Texas at Austin; Matthew Taylor, Washington State University; Lisa Torrey, St. Lawrence University; Shiqi Zhang, University of Texas at Austin.