Jochen Triesch studied physics and received his diploma degree in 1994 and his Ph.D. degree in 1999, both from the University of Bochum, Germany. After two years as a post-doctoral fellow at the University of Rochester, USA, he joined the faculty of the University of California San Diego in 2001 as an assistant professor of cognitive science. In 2005, he moved to the newly founded Frankfurt Institute for Advanced Studies (FIAS) in Frankfurt am Main, Germany. Since 2007 he is the Johanna Quandt Research Professor for Theoretical Life Sciences at FIAS. His research spans a number of different areas in different disciplines, including Neural Computation, Computational Neuroscience, Computer Vision, Machine Learning, Visual Psychophysics, Developmental Psychology, and Developmental Robotics.
Talk Title: Self-calibration of sensorimotor loops in active perception
Abstract: In many biological and artificial sensory systems, perception is an active process that involves movements of the sense organs such as the eyes, ears, or whiskers. These movements must be learned and calibrated. In biological systems this process is occurring autonomously and continuously. The mechanisms underlying such self-calibration are still largely unknown. We have recently proposed a new theoretical framework called Active Efficient Coding (AEC), that tries to explain how this self-calibration may take place. In a nutshell, AEC is a generalization of classic efficient coding theories in sensory neuroscience to active perception involving movements of the sense organs. AEC posits that the self-calibration of various sensorimotor loops results from the brain learning to control the movements of its sense organs to maximize its efficiency of encoding sensory inputs. We show the feasibility of this approach for the autonomous self-calibration of vergence and tracking eye movements and demonstrate how it can be successfully implemented on a humanoid robot.
Og DeSouza gained PhD in Ecology, Imperial College, UK. Associate Professor at Federal University of Viçosa, Brazil. Fellow of the Brazilian Council for Scientific Research. Member of the editorial board of the non-fee open source entomological journal "Sociobiology", the subscription-based entomological journal "Neotropical Entomology", and the non-fee open source publisher "CopIt ArXives". Researching on behavioural ecology of termites with focus on patterns emerging from individual behaviour and from interactions among individuals belonging to the same or to different species sharing the same nest.
Abstract: The living-together of distinct organisms in a single termite nest along with the termite builder colony, is emblematic in its ecological and evolutionary significance. On top of preserving biodiversity, these interspecific and intraspecific “symbioses” provide useful examples of interindividual associations thought to underly transitions in organic evolution. Being interindividual in nature, such processes may involve emergent phenomena and hence call for analytical solutions provided by computing tools and modelling, as opposed to classical biological methods of analysis. Here we provide selected examples of such solutions, showing that their use opens up wide and new research avenues in ecology and evolution.
Tony Pipe is Deputy Director of the Bristol Robotics Laboratory (http://www.brl.ac.uk/). He obtained his PhD qualification in 1997, became a Reader in 2006, and has been a full Professor of Robotics and Autonomous Systems since 2010. Tony Pipe has 20 years of experience in carrying out research on advanced sensor-systems, medical robotics, biologically-inspired robotics, machine learning and adaptive behaviour, applied to intelligent and distributed control/monitoring systems for robotics. His research foci are: innovative medical technology, safe physical Human Robot Interaction for robots co-located with humans; modelling animal brain signal processing and control structures; self-healing VLSI electronic hardware for safety-critical applications.
Talk Title: The rise of the machines?
Abstract: In the future, we may be able to receive a great deal of help from robots in handling many of today's "Dull, Dirty and Dangerous" tasks - the "3D"s'. 'Cognitively Capable Enough' robots could, perhaps, replace these tasks completely. However, there are both significant technical and societal issues to be tackled; we should all play a full part in making sure that we get the robot assistance that we really want, i.e., "Terminator oblivion" is not just around the corner.
- The 'assisted living' sector: a great deal can be done to allow us to live safely and independently for longer. Can this be achieved in technically, but crucially also a socially and psychologically acceptable way?
- Robots working as team members with humans in manufacturing environments: this may be a way to improve the working lives of many by allowing them to carry out some of the "3D's". That's good, but are there any down-sides?
- Military, security and surveillance: Robots may very useful, but their use is contentious. What are the issues?
In this talk, Prof. Pipe will use illustrative examples of research in BRL, and those of others, to talk a little more about some of the concepts introduced above, but he also wants to hear what you have to say on these topics!
Stefano Nolfi (http://laral.istc.cnr.it/nolfi/) is a research director of the Institute of Cognitive Sciences and Technologies of the Italian National Research Council and head of the Laboratory of Autonomous Robots and Artificial Life (http://laral.istc.cnr.it/). Stefano conducted pioneering research in Artificial Life and is one of the founders of Evolutionary Robotics. His main research interest is in study of how embodied and situated agents can develop behavioural and cognitive skills autonomously by adapting to their task/environment. Stefano authored and co-authored more than 150 peer-review scientific publications.
Talk Title: Evolution, Innovations, Evolvability
Abstract: Explaining how random genetic changes can generate complex useful innovations represents one of the hardest challenges for evolutionary theory. Progresses in evolutionary, molecular and developmental biology have clarified which are the processes and mechanisms that promote innovations and evolvability. However, we are not yet able to design artificial systems analogous to natural systems in that respect. In this lecture I will discuss the factors that could represent a pre-requisites for the occurrence of open-ended evolution in artificial system. Moreover I will describe the experimental evidences, collected by experimenting with artificial evolving systems, that demonstrate the importance and the role of these factors.