[robotics-worldwide] [meetings] CfP: IEEE RO-MAN 2015 workshop "memX: memory-based architectures for human-robot interaction"

Fulvio Mastrogiovanni fulvio.mastrogiovanni at unige.it
Mon Apr 13 13:25:20 PDT 2015


Dear All,

You are cordially invited to contribute and participate to memX, our workshop
co-located with IEEE RO-MAN 2015 in Kobe, Japan, on August 31 2015.

Cheers,

Fulvio

==========

memX: memory-inspired robot architectures for human-robot interaction

Website: https://memxws.wordpress.com/

*** Rationale ***

memX aims at investigating and fostering the discussion about the development of
memory-based architectures for robots, with a specific emphasis on their impact
on human-robot and robot-environment interaction.

During the past few years, many strands of activity have been put forward to
enhance the perceptual, cognitive, reasoning and action capabilities of robots.
On the one hand, the somewhat classical research activity on perception,
knowledge representation, and action planning and execution has reached a high
maturity level, which allows robots to operate in environments which do not
have to be structured for them beforehand, although with a few limitations. On
the other hand, it is now commonly accepted that novel forms of intuitive
human-robot interaction require robots to use multi-modal perceptual
information, efficient knowledge representation frameworks, common sense
reasoning systems, as well as advanced forms of action, including manipulation,
in a complete perception-action cycle.

We posit that memory-based robot architectures can act as a general framework to
ground different aspects of robot behaviours, and specifically a purposive
interaction with the environment, objects and humans therein.

Different application scenarios are envisaged. These include: companion robots,
which must understand and adapt to human habits, over (possibly) long periods
of time; co-workers, which must cooperate with human workers to aid them when
performing difficult tasks (e.g., assemblage); or other service robots, in
which the knowledge of (possibly complex) tasks to carry out requires a tight
integration among different components of the robot architecture.

Independently of the specific application scenario, at the scientific and
technological levels, three key research activities must be pursued to build
advanced robot frameworks inspired by memory processes.

The first involves how we actually design a robot architecture that mimics
relevant parts of what we currently understand be the relevant traits of
memory. This effort involves modelling sensory memory, short-term and long-term
memory, as well as relevant sub-components (e.g., procedural memory).
Furthermore, it implies research work on how memory is bootstrapped and which
role the robot body plays.

The second is related to the relationships between the memory model and the
perceptual, cognitive, reasoning and action behaviours of robots. These include
memory items formation, retreat, revision, and the use of assessed memory items
into knowledge, which can be used by robot behaviours in human-robot
interaction processes, also in the form of language.

The third is related to how we can make memory-based architectures efficient
from a computational standpoint. Robots operating in unstructured environments
and interacting with humans, must be able to take fast decisions about possible
courses of action. This means that specific requirements must be enforced in the
whole framework, possibly employing real-time or any-time approaches to
information processing.

memX will address these three aspects through focused sections.
Contributions are welcome, which address the following topics (and beyond):

*Memory-inspired robot architectures
*Memory modelling
*Memory-inspired knowledge representation
*Robot behaviour grounding on memory models and procedural knowledge
*Memory and multi-modal robot perception
*Memory items consolidation
*Retrieval of memory items for human-robot interaction
*Computational aspects of memory-based architectures
*Memory-based reasoning

*** Important Dates ***

* Two-page abstract deadline: June 30, 2015
* Notification of acceptance: July 10, 2015
* Camera-ready abstract deadline: July 20, 2015
* memX date: August 31, 2015

*** Submissions ***

Researchers wishing to present their ideas and results at memX are kindly
requested to submit a 2-page extended abstract, including figures, tables and
references.

Extended abstracts will form the Workshop Proceedings.

We have established preliminary contacts with publishers to evaluate a
post-Workshop extended publication of results, either as an edited book or a
journal special issue. More on this soon.

*** Organizers ***

Fulvio Mastrogiovanni
Department of Informatics, Bioengineering, Robotics and Systems Engineering,
University of Genoa, Italy
Email: fulvio.mastrogiovanni at unige.it

Nak Young Chong
Japan Advanced Institute of Science and Technology, Kanazawa, Japan
Email: nakyoung at jaist.ac.jp

Il Hong Suh
Hanyang University, Seoul, South Korea.
Email: ihsuh at hanyang.ac.kr

Hirokazu Tanaka
Japan Advanced Institute of Science and Technology, Kanazawa, Japan
Email: hirokazu at jaist.ac.jp

*** Program Committee ***

Yiannis Aloimonos
University of Maryland, USA

Tamim Asfour
Karlsruhe Institute of Technology, Germany

David Filliat
ENSTA-ParisTech, France

Nikolas Hemion
Aldebaran Robotics, France

Matthias Rolf
Osaka University, Japan

Gregor Schoener
Ruhr-Universitat Bochum, Germany

Alessandra Sciutti
Italian Institute of technology, Italy

Alexander Stoytchev
Iowa State University, USA

Jun Tani
KAIST, South Korea


More information about the robotics-worldwide mailing list