[robotics-worldwide] CfP to the IROS 2013 Workshop on Metrics of Embodied Learning processes in robots and animals

Fabio Bonsignorio fabio.bonsignorio at uc3m.es
Mon Jun 10 07:13:02 PDT 2013


Apologies for multiple postings
-------------------------------------------------------------------------------


Dear All,

you are kindly to submit to the IROS 2013
Workshop on Metrics of embodied learning processes in robots and animals on
November, 7th 2013

A key feature of biological intelligent systems is their ability to adapt
their behaviors to the changes of an open-ended environment. These
capabilities have so far only partially been mimicked by robot systems.
Learning processes are necessary to adapt to a stochastically-varying
environment and even to manage the robot itself if the robot has a soft,
rather than rigidly defined, structure. Moreover, in robotics (and in
nature) learning algorithms execute within an embodied agent: how does this
affect their performance? Should we explicitly consider the body morphology
‘*inside’* the learning algorithms?

How should one compare different implementations? How should we measure the
benefits of ‘embodiment’? of soft structure implementation? Of
bioinspiration? How can we compare our systems with biological examples?

It is worth mentioning, once again, that experimental methods and methods
to quantitatively evaluate the performance in similar classes of function
adopted in Robotics are sometimes coarser than those used by natural
sciences such as neurophysiology/neurosciences and biology.

The aim of this workshop is to discuss the fundamental issues raised by
performance measurement of cognitive and intelligent systems in the context
of a key class of problems: those raised by learning in embodied agents,
artificial and natural.

We think this will contribute to clarify and make progress in the robot
learning domain.

We aim to make another step towards a replicable robotics research
publishing thread based on the publication of fully replicable experiments,
which will reduce the distance from disciplines studying artificial and
natural autonomous systems.

The best contributions will be invited to submit to a refereed edited book.

All the best
Fabio, Angel, Elena and John


**Submission info:

Key Dates:

September 3rd, 2013 (23:59 PST): Paper Submission Deadline
September ,10th, 2013: Notification of Acceptance
October, 7th, 2013: Final version

Abstract Format:

Abstract should be submitted as PDF files (from minimum 1  to  maximum 2
pages, plus references)
in IROS regular paper format.

*List of topics:*

* Metrics of learning processes in robots and animals; *Replication of
learning process experiments in robots and animals; *Metrics of embodiment;
Metrics in Soft Robotics; *Replication of Soft Robotics experiment;
*Metrics of growing processes in robots and animals: *Replication of
growing process experiments in robots; Metrics for sensory motor
coordination and visual servoing effectiveness and efficiency; * Metrics
for shared control effectiveness and efficiency; *Benchmarking autonomy and
robustness to changes in the environment/task; *Scalable autonomy
measurements; *Design of Experiments in Robotics/AI; * Execution of
Experiments in Robotics; *Comparison of experimental methodology in
neurosciences/neurophysiology and in Robotics; *Comparison of experimental
methodology in neurosciences/neurophysiology and in AI; *Comparison of
experimental methodology in Biology and in Robotics; *Comparison of
experimental methodology in Biology and in AI; *Success metrics in
bioinspired Robotics; * Experimental scenarios to evaluate performance,
demonstrate generality, and measure robustness in robots and animals; *
Well grounded experimental methods to compare artificial and natural
cognitive systems; * Proposals for Promotion of Good Experimental Work;
Proposals for benchmark standardization of learning processes.


More information about the robotics-worldwide mailing list