ALTENBERG SEMINARS IN THEORETICAL BIOLOGY
Winter 2002/2003:
Evolutionary Robotics
The Topic
Hörsaal 2, Biozentrum, Althanstrasse 14,
at 6.15 p.m.
The Program at a glance:
Please note the change of dates for Kerstin Dautenhahn`s Talk!
12 December 2002
Barbara WEBB (Stirling):
"The Use of Robots to Explore Biological Hypotheses"
9 January 2003
Kerstin DAUTENHAHN (Hertfordshire):
"Social Intelligence and Imitation in Animals
and Robots"
30 January 2003
Dario FLOREANO (Lausanne):
"Addressing Biological Questions with Evolutionary
Robotics"
13 February 2003
Thomas CHRISTALLER (Sankt Augustin):
"Artificial Intelligence: What Does it Really
Mean (Today)"
20 February 2003
Inman HARVEY (Sussex):
"Evolutionary Robotics: A New Scientific Tool for
Studying Cognition"
The topic
Evolutionary Robotics
Research in the life sciences is increasingly data- and technology-driven. When trying to meet the challenge of articulating a Theoretical Biology, we must take into account the circumstance that evidence in biology is of a three-fold nature: in vivo, in vitro, and in silico — computer modeling and simulation. The methodological complexities we face are further increased by the fact that we are no longer only representing and intervening in living matter, but are now also creating it; science thus merges with technology..
The Winter 02/03 Altenberg Seminars in Theoretical Biology will focus on one prominent aspect of this transition: the development of Evolutionary Robotics (ER). What is at stake? What are its promises and limitations?
Living organizations are characterized, among other things, by learning (in-di-vidual and collective), adaptation to changing environments, metabolism, and selective reproduction. Complex, successful organisms can result from the synergistic operation of simple, automatic adaptive mechanisms. In the field of ER these concepts are applied to robot design. Rather than design robots whose hardware and software are fully specified, humans create robots with simple, adaptive mechanisms and selective reproductive capability.
Five prominent speakers will discuss the emerging conceptual founda-tions and methodological promises of the field, the use of robots to explore biologi-cal questions and hypo-theses concerning evolution and development, as well as social robotics. In addition, the potential of ER for tackling issues con-cerning the basics of cognition, per-cep-tion, and learning will be critically as-sessed.
Links:
http://www.evobot.com/
http://www.geocities.com/atreshan/links.html
Abstracts and biographical notes
Barbara WEBB
Department of Psychology
Stirling University, UK
Homepage
The Use of Robots to Explore Biological Hypotheses
Thursday 12 December
Abstract
'Evolutionary robotics' (ER) generally means using some form of random generation and selection to design or adapt a robotic system to its environment. In this talk I will consider the extent to which this approach has relevance to biological understanding. Have we learnt anything new about the nature of evolution from these robots? Or has the approach contributed to biology in other ways, for example, illuminating the function of some existing biological design by exploring the space of possible designs? The use of robots will be compared to simulation and mathematical models; and the evolutionary approach compared to more direct modelling of biological systems. In particular I will discuss how ER might be used to explore the contribution of sensory bias to sexual selection.
Biographical note
Barbara WEBB joined the Psychology Department at Stirling University
in 1999. She received her PhD in Artificial Intelligence from the University
of Edinburgh in 1993, and her BSc in Psychology from the University of
Sydney in 1988. She previously lectured at the universities of Edinburgh
(1993-1995) and Nottingham (1995-1998).
Dr. WEBB’s main research interest is in perceptual systems for the
control of behavior. This work is largely concerned with building computational
and physical models of these mechanisms to explicate and evaluate hypotheses.
In particular, she has developed and tested a robot model of the auditory
localization behavior of the cricket. Her current work is concerned with
integrating additional sensorimotor systems onto the same robot, to study
problems of interaction of basic behaviors. She also has an interest in
theoretical issues of methodology; in particular the problems of measurement,
modeling, and simulation.
Selected publications
WEBB, B./HARRISON, R. 2000. Eyes and ears: Combining sensory motor systems modelled on insect physiology. IEEE International Conference on Robotics and Automation.
WEBB, B./SCUTT, T. 2000. A simple latency dependent spiking neuron model of cricket phono-taxis. Biological Cybernetics 82:247—269.
WEBB, B. 1999. A framework for models of biological behaviour. International Journal of Neural Systems 9: 375—381.
WEBB, B. 1998. Robots crickets and ants: Models of neural control of chemotaxis and phono-taxis. Neural Networks 11: 1479—1496.
LUND, H.H./WEBB, B./HALLAM, J. 1998. Physical and temporal scaling considerations in a robot model of cricket calling song preference. Artificial Life 4: 95—107.
WEBB, B./HALLAM, J. 1996. How to attract females: Further robotic experiments in cricket phonotaxis. In From Animals to Animats 4: Proceedings of the 4th International Conference on the Simulation of Adaptive Behavior, 75—83. MIT Press.
WEBB, B. 1996. A robot cricket. Scientific American 275(6): 94—99.
Kerstin DAUTENHAHN
Adaptive Systems Research Group
University of Hertfordshire, UK
Homepage
Social Intelligence and Imitation in Animals and Robots
Thursday 9 January
Abstract
My talk will provide an overview on the field of social robotics. Inspired by studies on social behavior in primate and other individualized societies I defined social robots as follows (DAUTENHAHN 1999): embodied agents that are part of a heterogeneous group – a society of robots or humans; they are able to recognize each other and engage in social interactions; they possess histories (perceive and interpret the world in terms of their own experience), and they explicitly communicate with and learn from each other. I will give examples of research in this area and point out differences to the related field of collective robotics, where social behavior of robots is inspired by social insect societies. A particularly challenging research topic in social robotics (which is very relevant for robot-robot as well as robot-human interactions) is social learning and imitation. Artificial Intelligence researchers often consider imitation an effective mechanism by which robots can learn new skills. However, in humans and other animals imitation plays an important social role, too. My talk will give examples of research with such social robots that can learn from each other by imitation, and how this is related to research on imitation in biology and psychology.
Biographical note
Kerstin DAUTENHAHN is Reader in Artificial Intelligence in the Department
of Computer Science, Adaptive Systems Research Group, University of Hertfordshire.
She earned a doctoral degree at the University of Bielefeld, Department
of Biological Cybernetics, in 1993. She is a former member of the AI division
of the German National Research Center for Information Technology (GMD),
Germany, the AI Lab at Vrije Universiteit Brussel, and the Department
of Cybernetics at University of Reading in the UK.
Dr. DAUTENHAHN teaches courses on Artificial Life and behavior-based robotics.
She has published more than 100 papers in the areas of Socially Intelligent
Agents, Social Robotics, and Artificial Life. She edited 10 special journal
issues on related subjects in different international journals.
Selected publications
DAUTENHAHN, K./BOND, A./CAÑAMERO, L./EDMONDS, B., eds.. 2002. Socially Intelligent Agents: Creating Relationships with Computers and Robots. Kluwer Academic.
DAUTENHAHN, K./NEHANIV, C.L., eds. 2002. Imitation in Animals and Artifacts. MIT Press.
DAUTENHAHN, K. 1999. Embodiment and interaction in socially intelligent life-like agents. In NEHANIV, C.L., ed., Computation for Metaphors, Analogy and Agents. Springer Lecture Notes in Artificial Intelligence, Vol. 1562, 102—142.
DAUTENHAHN, K. 1995. Getting to know each other: Artificial social intelligence
for autonomous robots. Robotics and Autonomous Systems 16: 333—356.
Dario FLOREANO
Autonomous Systems Lab
Swiss Federal Institute of Technology
Lausanne (EPFL), Switzerland
Homepage
Addressing Biological Questions with Evolutionary Robotics
Thursday 30 January
Abstract
Evolutionary Robotics (ER) is a method for the automatic creation of
control systems and morphologies of embodied and situated robots. ER generates
control systems while the robot interacts with its environment. ER is
used both as a tool to investigate models and theories of biological life
and a method to generate powerful (and yet simple) control systems for
autonomous robots. After a short introduction to the methodology, I will
address the following open issues in biology by means of evolutionary
robots:
1) competitive co-evolution and progress;
2) functionality and communication in networks of spiking neurons;
3) co-evolution of active vision and feature selection in behavioral systems.
Biographical note
Dario FLOREANO is Professor of the Swiss National Science Foundation at the Autonomous System Laboratory, Swiss Federal Institute of Technology in Lausanne (EPFL). His areas of activity are ER, adaptive autonomous agents, bioinspired hardware, swarm intelligence, neural networks, and artificial life. He received a BA in Visual Psychophysics (1988) from the University of Trieste, an MS in Neural Computation (1992) from the University of Stirling, and a PhD in Cognitive Science and Robotics (1995) from the University of Trieste. He has occupied senior research positions at the National Research Council in Rome, the University of Stirling, the Swiss Federal Institute of Technology, and Sony Computer Science Laboratory in Tokyo. Prof. FLOREANO published more than 70 peer-reviewed papers.
Selected publications
FLOREANO, D. 2002. Ago ergo sum. In FETZER, J.H., ed., Consciousness Evolving, 181—204. John Benjamins.
FLOREANO, D./MATTIUSSI, C. 2002. Manuale sulle Reti Neurali. 2nd ed. Il Mulino.
HALLAM, J./FLOREANO, D./HAYES, G./MEYER, J.A., eds. 2002. From Animals to Animats 7. Proceedings of the 7th International Conference on Simulation of Adaptive Behavior. MIT Press.
FLOREANO, D./NOLFI, S./MONDADA, F. 2001. Co-evolution and ontogenetic change in competing robots. In PATEL, M./HONAVAR, V./BALAKRISHNAN, K., eds., Advances in the Evolutionary Synthesis of Intelligent Agents.
FLOREANO, D./URZELAI, J. 2000. Evolutionary robotics: The next generation. In GOMI, T., ed., Evolutionary Robotics III. AAI Books.
MEYER, J.A./BERTHOZ, A./FLOREANO, D./ROITBLAT, H./WILSON, S., eds. 2000. From Animals to Animats 6. Proceedings of the 6th International Conference on Simulation of Adaptive Behavior. MIT Press.
NOLFI, S./FLOREANO, D. 2000. Evolutionary Robotics: The Biology, Intelligence, and Technology of Self-organizing Machines. MIT Press.
FLOREANO, D./, NICOUD, J.D./MONDADA, F., eds. 1999. Advances in Artificial Life. Proceedings of the 5th European Conference on Artificial Life. Springer.
Thomas CHRISTALLER
Fraunhofer-Institute for Autonomous intelligent Systems (AiS)
Schloss Birlinghoven
53754 Sankt Augustin, Germany
Homepage
Artificial Intelligence: What Does it Really Mean (Today)?
Thursday 13 February
Abstract
Central to the mainstream of AI research is the Physical Symbol System
Hypothesis of Newell and Simon. This hypothesis can be seen to be within
the long tradition of mathematical logic, in which the notion of truth
and the description of domains using logical formulas are important elements.
But this is reminiscent of the proverb which says that for someone with
a hammer the whole world is a nail. We need to ask: What is natural intelligence,
cognition or consciousness really?
With all caution which one must have as a scientist, most empirical findings
and biological theories can be interpreted in such a way that the symbolic
and logic-oriented approach falls short in substantial aspects, if the
issue is how to "equip" artifacts with intelligence, cognition,
and consciousness.
My main hypothesis is that intelligence enables organisms to make predictions
about the future, especially about the behavior of other members of their
own species. In my talk, I will present some central arguments supporting
this hypothesis.
Finally, emanating from this notion of intelligence, I will present some
work from the cognitive robotics group of my institute.
Biographical note
Thomas CHRISTALLER has been the Director of the Fraunhofer Institute
for Autonomous intelligent Systems (AiS) since 2001; he joined their research
group on Expert Systems in 1985.
Prof. CHRISTALLER studied mathematics, physics, and computer science at
the universities of Marburg and Bonn (1970-1977). He holds a PhD from
the University of Hamburg ("Generische Kontrollstrukturen am Beispiel
von kaskadierenden ATNs", 1986). He has previously worked at the
University of Bielefeld (1977-1981) and the University of Hamburg (1982-1984).
He has been full professor of Artificial Intelligence at the University
of Bielefeld since 1991. From 1991 to 1997 he has been the Director of
the GMD Institute of Applied Information Techniques, and from 1998 to
2001 the Director of the GMD Institute of System Design Technology (called
Institute for Autonomous intelligent Systems since 1999).
Selected publications
CHRISTALLER, TH./PAETAU, M. 2001. Informed sustainability: Autonomy and knowledge for sustainable development.
KNOLL, A./CHRISTALLER, TH. 2000.
Selbstrepräsentation, Selbstwahrnehmung und Verhaltenssteuerung von
Robotern. In SANDKÜHLER, H.-J., Hg., Selbstrepräsentation
in Natur und Kultur, 109—132.
KOLO, C./CHRISTALLER, TH./PÖPPEL, E. 1999. Bioinformation (Problemlösungen für die Wissensgesellschaft). Physica-Verlag.
HERTZBERG, J./CHRISTALLER, TH./KIRCHNER, F./LICHT, U./ ROME, E. 1998.
Sewer robotics. In PFEIFER, R., ed., From Animals to Animats 5, 427—436.
DAUTENHAHN, K./CHRISTALLER, TH. 1997. Remembering,
rehearsal and empathy: Towards a social and embodied cognitive psychology
for artifacts. In NUALLAIN, O.S./MCKEVITT, P./MAC AOGAIN, E., eds.,
Two Sciences of Mind.
SCHLOTTMANN, E./SPENNEBERG, D./PAUER, M./CHRISTALLER, TH./DAUTENHAHN,
K. 1997. A
modular design approach towards behaviour oriented robotics. Arbeitspapiere
der GMD, 1088, GMD, Sankt Augustin.
MAAR, C./PÖPPEL, E./CHRISTALLER, TH. 1996. Die Technik auf dem Weg zur Seele. Forschungen an der Schnittstelle Gehirn/Computer. Rowohlt.
CHRISTALLER, TH. 1995. Probehandeln
als ein imaginativer Prozeß. In Illusion und Simulation –
Begegnung mit der Realität, 158—174. Cantz.
top
Inman HARVEY
Centre for Computational Neuroscience and Robotics &
Evolutionary and Adaptive Systems Group
School of Cognitive and Computing Sciences
University of Sussex, Falmer, Brighton
Homepage
Evolutionary Robotics:
A New Scientific Tool for Studying Cognition
Thursday 20 February
Abstract
Evolutionary Robotics allows the experimenter to try and re-create some
cognitive phenomena of interest under controlled conditions, whilst minimizing
the preconceptions and prejudices that are put into the model. In this
way ER is potentially a new scientific tool for studying the basics of
cognition, perception, and learning. A thought experiment can be tested
on either real or simulated robots/agents, typically giving artificial
evolution a free hand to design a real-time artificial neural network
that generates behavior selected to demonstrate the desired properties.
Some examples from within our research group at Sussex include:
• the origins of basic visual perception and object recognition;
• homeostasis and the ability of agents to overcome severe sensory
distortion;
• the origins of communication;
• the circumstances under which learning becomes important; and
• the interactions between evolution and development.
As these experiments are done in a very abstract setting, the implications
can be very general and widespread.
Biographical note
Inman HARVEY is Senior Research Fellow at the CCNR (Centre for Computational
Neuroscience and Robotics) as well as a member of the EASy (Evolutionary
and Adaptive Systems) Group at the University of Sussex’ COGS (School
of Cognitive and Computing Sciences), which he joined 13 years ago and
helped to found while pursuing a doctorate in the development of artificial
evolution for design problems. He has a background in mathematics, philosophy,
anthropology, and… oriental carpet-dealing. He is interested in the
development of artificial evolution as an approach to the design of complex
systems (e.g., evolutionary robotics, evolvable hardware, molecules for
pharmaceutical purposes).
A current focus is Neutral Networks — pathways of neutral mutations
through sequence space, or percolating `ridges' through fitness landscapes,
which can be exploited by artificial evolution. Evolutionary robotics
links artificial evolution to Prof. HARVEY’s philosophical approaches
to Artificial Life. He describes himself as a relativist who rejects classical
AI, and treats cognition as the outcome of sensori-motor couplings of
an agent with its world. His “pet hates” include “unthinking
closet Platonists, careless and slip-shod use of words such as `internal
representation', `learning', `information', and particularly `computation'.”
Prof. HARVEY has recently been much involved with a Sci-Art project to
build a massive 6-legged man-carrying walking robot for use in performance
art, and to demonstrate dynamical approaches to walking.
Selected publications
HARVEY, I. 2002. Evolving robot consciousness: The easy problems and the rest. In FETZER, J.H., ed., Evolving Consciousness, 205—219. John Benjamins.
SHIPMAN, R./SHACKLETON, M./HARVEY, I. 2000. The use of neutral genotype-phenotype mappings for improved evolutionary search. BT Technology Journal 18(4): 103—111.
HARVEY, I. 2000. Robotics: Philosophy of mind using a screwdriver. In GOMI, T., ed., Evolutionary Robotics: From Intelligent Robots to Artificial Life, Vol. III, 207—230. AAI Books.
HUSBANDS, P./HARVEY, I./CLIFF, D./MILLER, G. 1997. Artificial evolution: A new path for artificial intelligence? Brain and Cognition 34: 130—159.
HARVEY, I. 1997. Cognition is not computation: evolution is not optimisation. In GERSTNER, W./GERMOND, A./HASLER, M,/NICOUD, J.-D., eds., Artificial Neural Networks — ICANN97, 685— 690. Springer.
HARVEY, I. 1997. Is there another new factor in evolution? Evolutionary
Computation (Special Issue on Evolution, Learning, and Instinct: 100 Years
of the Baldwin Effect) 4(3): 311— 327.
CLIFF, D./HARVEY, I./HUSBANDS, P. 1993. Explorations in evolutionary robotics.
Adaptive Behavior 2: 71—104.
HUSBANDS, P./HARVEY, I. 1992. Evolution versus design: Controlling autonomous
robots. In Integrating Perception, Planning and Action: Proceedings of
3rd Annual Conference on Artificial Intelligence, Simulation and Planning,
139—146. IEEE Press.