- MOTIVATION AND OBJECTIVES
The Workshop on Neurorobotics - New perspectives in the synergy between Neuroscience and Robotics will be organized on June 26, at the sixth IEEE RAS/EMBS International Conference on Biomedical Robotics and Biomechatronics – BioRob 2016, that will be held in Singapore.
Neurorobotics represents the two-front approach to the study of brain: Neuroscience attempts to elucidate brain mechanisms by investigating intelligent biological systems, while Robotics attempts to recreate sensory-motor coordination by building bio-inspired machines, with control based on machine learning and other bio-inspired computing techniques.
Neurorobotics is already an established branch of robotics that in the last years, taking advantage from an increase in the accessibility of existing neuroscientific data and knowledge, allowed building robotic systems that can exhibit robustness, adaptability and several features of the human intelligence. Reciprocally, significant developments in robotics and machine learning put robotics in the service of neuroscience as experimental platforms or test-beds of brain models.
In the last years, advanced insights and the increasing availability of cheap processing power has led Neurorobotics to follow two tracks of research with different goals and methods:
The first track focuses on neuro-inspired computing paradigms that mimic nervous system functions based on Spiking Neural Networks. This does not only foster our understanding of biological systems but also contributes to future technical applications in artificial systems. In the past, limited processing power and the lack of appropriate models and tools shifted the focus of research far away from biological neural networks. Recently a number of projects like the US BRAIN Initiative and the Human Brain Project have taken up the challenge by combining efforts from the fields of neuroscience and computer science to enable the large scale modeling and simulation of biological neural networks with billions of spiking neurons.
The second track, extending the theory of classical artificial neural networks, mostly relies on simpler neuron models but integrate them in novel network architectures. These networks are extensively used in robotics, allowing mimicking the function of some brain areas in order to reproduce complex behaviors with a reduced computational cost.
This workshop seeks to present and discuss advances in neuroscientific models for cognition and new perspectives in control for robotic applications based on both biologically-inspired and artificial neural networks. The final goal is to bring together researchers from both robotics and neuroscience in order to explore how to maximize the progress at the multidisciplinary frontier evaluating the advantages of both tracks of the Neurorobotics research.
TOPICS OF INTEREST
¦ Bio-inspired sensory-motor coordination and adaptive control
¦ Active perception
¦ Self-organization and sensory-motor mapping
¦ Predictive behaviour
¦ Bio-inspired learning robots
¦ Robot imitation and learning by demonstration
¦ Memory-based algorithms
¦ Cognitive behaviours in robots
¦ Reservoir computing
¦ Deep learning
¦ Neuromorphic computing for robotics
June 26, 2016 (tentative agenda)
- 14.30 - 14.40: Florian Röhrbein [Technical University of Munich]
Welcome and Introduction to the Human Brain Project
- 14.40 - 15.00: Egidio Falotico [Scuola Superiore Sant’Anna]
“Connecting artificial brains to robots in a comprehensive simulation framework: the Neurorobotics Platform”
- 15.00 - 15.30: Giorgio Metta [Istituto Italiano di Tecnologia, Italy]
“The use of motor invariants can improve action discrimination”
- 15.30 - 15.45: Coffee Break
- 15.45 - 16.15: Tomohiro Shibata [Kyushu Institute of Technology, Japan]
“Functional Robotic Assistance of Human Motor Learning
- 16.15 - 16.45: Ruediger Dillman [Research Center for Information Technology, Germany]
“Modelling Cortical Sensor-Motor Control Functionalities with a Spiking Neural-Robot Control Simulato”
- 16.45 - 17.15: Satoshi Oota [Riken Bioresource Center, Japan]
“A study on mouse locomotor functions with a disturbance-based approach”
- 1715 - 17.30: Florian Röhrbein [Technical University of Munich]
Conclusions and final discussion
Giorgio Metta [Istituto Italiano di Tecnologia, Italy] - “The use of motor invariants can improve action discrimination”
Tomohiro Shibata [Kyushu Institute of Technology, Japan] - “Functional Robotic Assistance of Human Motor Learning”
Ruediger Dillman [Research Center for Information Technology, Germany] - “Modelling Cortical Sensor-Motor Control Functionalities with a Spiking Neural-Robot Control Simulator” Abstract
Satoshi Oota [Riken Bioresource Center, Japan] - “A study on mouse locomotor functions with a disturbance-based approach” Abstract
|University Town / UTown
1 Create Way
Create Tower Level 1, UTown
• Call for Contributions:
April 14, 2016 (extended)
• Notification of acceptance:
April 21, 2016