CFP Special track on "Bio-inspired error detection" (BIONETICS 2011)

December 5-7, 2011, York, UK

http://www.bionetics.org/


TRACK CHAIRS: Iain Bate (University of York, UK) and Martin Drozda
(Leibniz University of Hannover, Germany)

SCOPE AND TOPICS:

One of the most important properties of biological systems is their
capability to detect changes in external environment, to adapt to these
changes and to form self-maintaining systems with internal feedback
paths allowing for reactions to these changes. These capabilities of
biological systems remain unrivaled, despite huge research effort and
expense committed to the design of their human-made counterparts.

Errors can emerge as a consequence of poor system design,
software&hardware faults or malicious behavior. The ambition of
human-designed error detection systems is to provide several key
properties including low classification error, ability to detect novel
error types, ability to adapt with the goal to provide an improved
detection performance upon error re-occurrence and ability to provide
initial calibration so that a large set of well described errors can be
detected immediately upon the deployment of an error detection system.
Objectives such as energy efficiency may also become an integral design
requirement of error detection systems.

This special session aims at bringing together contributions that apply
ideas inspired by natural systems to error detection systems.
It will address leading edge research ranging from theory to
experimental evaluation, tools and case studies. The bio-inspired topics
of interest include but are not limited to:

o Architectures and models for error detection including architectures
and models based on cascading or voting classification, and including
multi-component architectures inspired by the Biological immune system
o Adaptive error detection including feature identification and
extraction methods that facilitate efficient bio-inspired adaptive error
detection
o Novel error detection with focus on bio-inspired two-class classification
o Feed-forward/feed-back error detection
o Energy efficient error detection including approaches based on
heterogeneous agents, each exhibiting a unique detection capability and
energy cost
o Active reaction to errors with focus on reactions that help
re-introduce a stable-state operation
o Error detection approaches for enhanced dependability and
survivability
o Techniques for initial or sporadic error detection system calibration
such as vaccination
o Performance evaluation of error detection systems, especially when one
or several bio-inspired components interact with other machine learning
approaches and/or statistical methods


Important Dates:

Paper submission due: September 1st, 2011
Notification of acceptance: September 23rd, 2011
Camera ready due: October 8th, 2011

Paper Submission:

Authors are invited to submit papers of up to 15 pages in the Springer
LNICST format. Please visit http://www.bionetics.org/submission.shtml
for detailed submission instructions.