AIM AND SCOPE

Although machine learning algorithms continue to improve at a rapid pace enabling technologies and products such as autonomous driving cars and sophisticated image and speech recognition, it is often forgotten that these applications represent tailored solutions to specific tasks. Thus it is not clear if or how these autonomous systems can pave the road to general purpose machines envisioned by many.

The pursuit for higher levels of autonomy and versatility in robotics is arguably lead by two main factors. Firstly, as we push robots out of the labs and productions lines, it becomes increasingly difficult to design for all possible scenarios that a particular robot might encounter. Secondly, the cost of designing, manufacturing, and maintaining such systems becomes prohibitive.

As the algorithms for learning single tasks in restricted environments are improving, new challenges have gained relevance in order to get more autonomous artificial systems. These challenges include multi-task learning, multimodal sensorimotor learning and lifelong adaptation to injury, growth and ageing. Addressing these challenges promise higher levels of autonomy and versatility of future robots.

This special issue on Continual Unsupervised Sensorimotor Learning is primarily concerned with the developmental processes involved in unsupervised sensorimotor learning in a life-long perspective, and in particular the emergence of representations of action and perception in humans and artificial agents in continual learning. These processes include action-perception cycle, active perception, continual sensory-motor learning, environmental-driven scaffolding, and intrinsic motivation.

The special issue will highlight behavioural and neural data, and cognitive and developmental approaches to research in the areas of robotics, computer science, psychology, neuroscience, etc. Contributions might focus on mathematical and computational models to improve robot performance and/or attempt to unveil the underlying mechanisms that lead to continual adaptation to changing environment or embodiment and continual learning in open-ended environments.

Contributions from multiple disciplines including cognitive systems, cognitive robotics, developmental and epigenetic robotics, autonomous and evolutionary robotics, social structures, multi-agent and artificial life systems, computational neuroscience, and developmental psychology, on theoretical, computational, application-oriented, and experimental studies as well as reviews in these areas are welcome.

THEMES

This special issue aims to report state-of-the-art approaches and recent advances on Continual Unsupervised Sensorimotor Learning with a cross-disciplinary perspective. Topics relevant to this special issue include but are not limited to:

• Emergence of representations via continual interaction
• Continual sensory-motor learning
• Action-perception cycle
• Active perception
• Environmental-driven scaffolding
• Intrinsic motivation
• Neural substrates, neural circuits and neural plasticity
• Human and animal behaviour experiments and models
• Reinforcement learning and deep reinforcement learning for life-long learning
• Multisensory robot learning
• Multimodal sensorimotor learning
• Affordance learning
• Prediction learning

SUBMISSION

Manuscripts should be prepared according to the "Information for Authors" of the journal found at https://cis.ieee.org/publications/t-cognitive-and-developmental-systems/tcds-information-for-authors. Submissions must be done through the IEEE TCDS Manuscript center: https://mc.manuscriptcentral.com/tcds-ieee. Please select the category “SI: Continual Unsupervised Sensorimotor Learning”.

IMPORTANT DATES 

GUEST EDITORS

Nicolás Navarro-Guerrero
Aarhus University, Aarhus, Denmark
nng@eng.au.dk

Sao Mai Nguyen
IMT Atlantique, France
nguyensmai@gmail.com

Erhan Öztop
Özyeğin University, Turkey
erhan.oztop@ozyegin.edu.tr

Junpei Zhong
National Institute of Advanced Industrial Science and Technology (AIST), Japan
joni.zhong@aist.go.jp