Dagstuhl Seminar 14081
Robots Learning from Experiences
( Feb 16 – Feb 21, 2014 )
Permalink
Organizers
- Anthony Cohn (University of Leeds, GB)
- Bernd Neumann (Universität Hamburg, DE)
- Alessandro Saffiotti (University of Örebro, SE)
- Markus Vincze (TU Wien, AT)
Contact
- Susanne Bach-Bernhard (for administrative matters)
Dagstuhl Seminar Wiki
- Dagstuhl Seminar Wiki (Use personal credentials as created in DOOR to log in)
Schedule
The ability to exploit experiences is an important asset of intelligent beings. Experiences provide a rich resource for learning, solving problems, avoiding difficulties, predicting the effects of activities, and obtaining commonsense insights. Current robots do not in general possess this ability, and this is a decisive reason for the often perceived "lack of intelligence" of current robotic systems: they repeat mistakes, do not learn to anticipate happenings in their environment, and need detailed instructions for each specific task. Consider an everyday task of a service robot, such as grasping a cup from a cupboard and bringing it to a person sitting at a table. This task may occur in many variations and under unpredictable circumstances. For example, persons may sit at different sides of a table, a direct path to the table may be blocked, the table may be cluttered with various objects, hot water may be ready or not, the cup on the shelf may be upside-down, etc. It is clearly unfeasible to provide the robot with precise instructions for all contingencies at design time or to specify tasks with highly detailed instructions for each particular concrete situation which may arise. Hence without such knowledge, robot behaviour is bound to lack robustness if the robot cannot autonomously adapt to new situations.
How would the robot, for example, avoid pouring coffee into an upside-down cup? Based on experiences with multiple pouring actions, the robot will have formed a conceptualisation of all concomitant circumstances of successful pouring, for example to pour into a "container". The robot may not know the name of this conceptualisation but will know that it must be open on top, hollow, empty, etc. Similarly, the robot may have encountered upside-down objects before and hence be able to conceptualise the corrective action of turning an object to make it a usable container.
In this seminar, we want to bring together experts and scholars from the robotics, learning, and knowledge representation communities to discuss current approaches to make robots learn from experiences. Emphasis will be on the representation of real-world experiences and on exploiting experiences for autonomous acting in a changing or partially unknown environment. Hopefully, this will be a step toward the creation of the next generation of cognitive robotic systems. The programme of the seminar will consist of a selection of talks offered by the invited participants, based on abstracts which will be solicited, and on invited surveys addressing the seminar topic from different perspectives. There will be sufficient time after each talk for discussions. In addition, there will be dedicated group discussion sessions, as well as informal discussion sessions in the evenings on topics arising during the seminar. A seminar proceedings will be published summarizing the seminar.
The ability to exploit experiences is an important asset of intelligent beings. Experiences provide a rich resource for learning, solving problems, avoiding difficulties, predicting the effects of activities, and obtaining commonsense insights. Current robots do not in general possess this ability, and this is a decisive reason for the often perceived "lack of intelligence" of current robotic systems: they repeat mistakes, do not learn to anticipate happenings in their environment, and need detailed instructions for each specific task.
Consider an everyday task of a service robot, such as grasping a cup from a cupboard and bringing it to a person sitting at a table. This task may occur in many variations and under unpredictable circumstances. For example, persons may sit at different sides of a table, a direct path to the table may be blocked, the table may be cluttered with various objects, hot water may be ready or not, the cup on the shelf may be upside-down, etc. It is clearly infeasible to provide the robot with precise instructions for all contingencies at design time or to specify tasks with highly detailed instructions for each particular concrete situation which may arise. Hence without such knowledge, robot behaviour is bound to lack robustness if the robot cannot autonomously adapt to new situations.
How would the robot, for example, avoid pouring coffee into an upside-down cup? Based on experiences with multiple pouring actions, the robot will have formed a conceptualisation of all concomitant circumstances of successful pouring, for example to pour into a "container". The robot may not know the name of this conceptualisation but will know that it must be open on top, hollow, empty, etc. Similarly, the robot may have encountered upside-down objects before and hence be able to conceptualise the corrective action of turning an object to make it a usable container.
This seminar has brought together experts and scholars from the robotics, learning, and knowledge representation communities to discuss current approaches to make robots learn from experiences. Emphasis was on the representation of real-world experiences and on exploiting experiences for autonomous acting in a changing or partially unknown environment.
- Michael Beetz (Universität Bremen, DE) [dblp]
- Sven Behnke (Universität Bonn, DE) [dblp]
- Alexandre Bernardino (Technical University - Lisboa, PT) [dblp]
- Mustafa Blerim (The American University of Paris, FR) [dblp]
- Richard Bowden (University of Surrey, GB) [dblp]
- Ivan Bratko (University of Ljubljana, SI) [dblp]
- Francois Bremond (INRIA Sophia Antipolis - Méditerranée, FR) [dblp]
- Anthony Cohn (University of Leeds, GB) [dblp]
- Luc De Raedt (KU Leuven, BE) [dblp]
- Krishna Sandeep Reddy Dubba (University of Leeds, GB) [dblp]
- Martin Günther (Universität Osnabrück, DE) [dblp]
- Manfred Hild (HU Berlin, DE) [dblp]
- Vaclav Hlavac (Czech Technical University, CZ) [dblp]
- David C. Hogg (University of Leeds, GB) [dblp]
- Lothar Hotz (Universität Hamburg, DE) [dblp]
- Lorenzo Jamone (Technical University - Lisboa, PT) [dblp]
- Stefan Konecny (University of Örebro, SE) [dblp]
- Marek S. Kopicki (University of Birmingham, GB) [dblp]
- Jos Lehmann (Universität Hamburg, DE) [dblp]
- Ales Leonardis (University of Birmingham, GB) [dblp]
- Masoumeh Mansouri (University of Örebro, SE) [dblp]
- Ralf Moeller (TU Hamburg-Harburg, DE) [dblp]
- Bernd Neumann (Universität Hamburg, DE) [dblp]
- Davide Nitti (KU Leuven, BE) [dblp]
- Laurent Orseau (AgroParisTech - Paris, FR) [dblp]
- Pierre-Yves Oudeyer (INRIA - Bordeaux, FR) [dblp]
- Federico Pecora (University of Örebro, SE) [dblp]
- Sebastian Rockel (Universität Hamburg, DE) [dblp]
- Alessandro Saffiotti (University of Örebro, SE) [dblp]
- Luis Seabra Lopes (University of Aveiro, PT) [dblp]
- Muralikrishna Sridhar (University of Leeds, GB) [dblp]
- Luc Steels (Free University of Brussels, BE) [dblp]
- Sebastian Stock (Universität Osnabrück, DE) [dblp]
- Georgi Stojanov (The American University of Paris, FR) [dblp]
- Aryana Tavanai (University of Leeds, GB) [dblp]
- Carme Torras (UPC - Barcelona, ES) [dblp]
- Emre Ugur (Universität Innsbruck, AT) [dblp]
- Markus Vincze (TU Wien, AT) [dblp]
- Jure Zabkar (University of Ljubljana, SI) [dblp]
- Jianwei Zhang (Universität Hamburg, DE) [dblp]
- Michael Zillich (TU Wien, AT) [dblp]
Classification
- artificial intelligence / robotics
Keywords
- Learning
- experiences
- cognitive systems