IEEE Computational Intelligence Society (CIS) Technical Committee on Cognitive and Developmental Systems (CDS)

Aim and Scope

The objective of this task force is to strengthen the links between evolutionary approaches in artificial intelligence and the developmental and cognitive systems community. There are many potential convergences and common discussion topics, including, but not limited to:

- A large part of the evolutionary intelligence community work on “Generative and Developmental Encoding” to evolve large, life-like neural networks (“artificial brains”); as the name suggest it, these encoding exploit developmental rules to make large neural networks that exhibit some features of biological nervous systems like modularity, hierarchy, repetition, symmetry, etc.

- A new trend in evolutionary intelligence is to study alternative selective pressures and, in particular, selective pressures that encourage exploration (novelty search, behavioral diversity, etc.); these techniques share many objectives and ideas with the exploration algorithms studied in the developmental robotics community (e.g. babbling).

- Evolution can be an effective metaphore to explain and therefore imitate the plasticity of the brain (see neural darwinism and neural replicators).

- Evolution and life-time learning are two complementary processes that interact at different scales and a key research avenue is to understand what should be designed by evolution and what should be left for life-time learning.

Task Force Chair:

Jim Torresen, University of Oslo, Norway

Task Force Vice-Chairs:

Gusz Eiben, Vrije Universiteit Amsterdam, Netherlands

Stéphane Doncieux, Sorbonne Université, France

Task Force Members:

Andrea Soltoggio, Loughborough University, UK

Anthony Clark, Missouri State University, US

Jared M. Moore, Grand Valley State University, US

Jean-Baptiste Mouret, INRIA, France

Jeff Clune, University of Wyoming/Uber AI Labs, US

Kyrre Glette, University of Oslo, Norway

Nicolas Bredeche, Sorbonne Université, France

Richard Duro, Universidade da Coruna, Spain

Sebastian Risi, IT University of Copenhagen, Denmark

Stefano Nolfi, National Research Council (CNR), Italy

Stefano Nichele, Oslo Metropolitan University, Norway

Future Activities

2019

• Stéphane Doncieux, lectures a course at the International Summer School on Intrinsically Motivated Open-Ended Learning (http://www.goal-robots.eu/announcements/events/summer-school-2019-2
• Stéphane Doncieux, "Open-ended learning in robotics", talk at the IMOL 2019 workshop (http://www.imol-conf.org
• 9th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics. August 19-22, 2019 in Oslo, Norway. Paper submission deadline 22 February, 2019. Early registration deadline June 15, 2019.
• Stéphane Doncieux and Sebastian Risi will co-chair the complex systems track at GECCO 2019.
• Nicolas Bredeche, Jean-Baptiste Mouret and Stéphane Doncieu will give a tutorial on ER at GECCO 2019.

Past Activities

2019

• Gusz Eiben and Kyrre Glette are robotics area chairs for EvoApplications 2019, http://www.evostar.org/2019/cfp_evoapps.php

2018

• Alex Alcocer (OsloMet) chaired the Autonomy Day 2018 – 2nd Workshop on Autonomous and Adaptive Systems, 3 May in Oslo, Norway including talks on ER.
• Stéphane Doncieux, Joshua Auerbach, Richard Duro and Harold P. de Vladar organized the workshop "Evolution in Cognition" at the GECCO conference in 2018, see https://sites.google.com/champlain.edu/eic2018/ .
• Stéphane Doncieux organised a workshop on cognitive and developmental robotics at Sorbonne University, the 19th of April 2018 (with Verena Hafner, Jose Santos Victor, Gianluca Baldassare, Emre Ugur, Raja Chatila and Doncieux as speakers)
• Stéphane Doncieux has also organized a research topic in Frontiers, see https://www.frontiersin.org/research-topics/4233/evolvability-environments-embodiment-emergence-in-robotics.
• Stéphane Doncieux organised an exhibition at the Palais de la Découverte from March to May 2018 (the link with some informations about that in French: http://www.palais-decouverte.fr/fr/ressources/docs-1chercheur1manip/des-robots-qui-apprennent/).
• July 2018: Tutorial on Evolutionary Robotics at GECCO 2018 N. Bredeche, S. Doncieux and J-B. Mouret.
• Stéphane Doncieux has given several talks about the DREAM project in different contexts (7 in 2018 for academics and 4 for the general public, with an audience varying between 20 people and more than 300).
• Kyrre Glette co-chaired the EvoROBOT track at EvoApplications 2018, http://www.evostar.org/2018/cfp_evorobot.php
• Video showing the DyRET robot at the University of Oslo and some outdoor experiments. Nbcnews.com, 25 June 2018
• Interview Tønnes Nygaard at the University of Oslo about the DyRET robot and evolutionary robotics. Digitaltrends.com, 23. may 2018
• Interview with Tønnes Nygaard about the DyRET robot and work developing it at the University of Oslo. Wired.com, 18. may 2018

Published papers

• McAllister R, Kahn G, Clune J, Levine S (2019) Robustness to Out-of-Distribution Inputs via Task-Aware Generative Uncertainty. International Conference on Robotics and Automation (ICRA). (pdf)
• Stanley KO, Clune J, Lehman J, Miikkulainen R (2019) Designing Neural Networks through Neuroevolution. Nature Machine Intelligence, 1:1, 24-35. (pdf)
• Miconi T, Rawal A, Clune J, Stanley KO (2019) Backpropamine: training self-modifying neural networks with differentiable neuromodulated plasticity. International Conference on Learning Systems (ICLR) (31% acceptance rate).
• Doncieux, S., and Laflaquiere, A. and Coninx, A. (2019) Novelty Search: a Theoretical Perspective, to appear in GECCO-2019 proceedings.
• "Open-Ended Learning: A Conceptual Framework Based on Representational Redescription": https://loop.frontiersin.org/publications/51950747, A paper from the DREAM project consortium about a formalisation of open-ended learning (in which evolutionary approaches are relevant)
• "Bootstrapping Q-Learning for robotics from Neuro-Evolution Results": https://ieeexplore.ieee.org/abstract/document/7879193, This is one instance of the formalisation described in the paper above: The authors start with evolution to generate examples out of which we can extract a representation that will allow to generalize and learn faster.
• Kenneth O. Stanley, Jeff Clune, Welcoming the Era of Deep Neuroevolution
• Tønnes F. Nygaard, Charles P. Martin, Eivind Samuelsen, Jim Torresen, Kyrre Glette.Real-World Evolution Adapts Robot Morphology and Control to Hardware Limitations. In GECCO18 (ACM), 2018.
• Tønnes F. Nygaard, Charles P. Martin, Jim Torresen, Kyrre Glette. Self-Modifying Morphology Experiments with DyRET: Dynamic Robot for Embodied Testing. Preprint: Arxiv, 2018.
• Tønnes F. Nygaard, Charles P. Martin, Jim Torresen, Kyrre Glette. Exploring Mechanically Self-Reconfiguring Robots for Autonomous Design. In Workshop on Autonomous Robot Design at ICRA18, 2018
• Soltoggio, A, Stanley, K. O., Risi, S., Born to Learn: the Inspiration, Progress, and Future of Evolved Plastic Neural Networks. Neural Networks (2018) DOI: 10.1016/j.neunet.2018.07.013 arXiv: https://arxiv.org/abs/1703.10371
• Stanton C, Clune J (2018) Deep curiosity search: Intra-life exploration improves performance on challenging deep reinforcement problems. NIPS Deep Reinforcement Learning Workshop. (pdf)
• Lehman J, Chen J, Clune J, Stanley KO (2018) Safe mutations for deep and recurrent neural networks through output gradients. Proceedings of the Genetic and Evolutionary Computation Conference (GECCO). (pdf)
• Conti E, Madhavan V, Such FP, Lehman J, Stanley KO, Clune J (2018) Improving exploration in evolution strategies for deep reinforcement learning via a population of novelty-seeking agents. Advances in Neural Information Processing Systems (NIPS) 32 (20% acceptance rate). (pdf)
• Lehman J, Chen J, Clune J, Stanley KO (2018) ES is more than just a traditional finite-difference approximator. Proceedings of the Genetic and Evolutionary Computation Conference (GECCO). (pdf)
• Miconi T, Clune J, Stanley KO (2018) Differentiable plasticity: Training plastic neural networks with backpropagation. ICML. (pdf)
• Such FP, Madhavan V, Conti E, Lehman J, Stanley KO, Clune J (2018) Deep neuroevolution: Genetic algorithms are a competitive alternative for training deep neural networks for reinforcement learning. NIPS Deep Reinforcement Learning Workshop. (pdf)
• Norouzzadeh MS, Nguyen A, Kosmala M, Swanson A, Palmer M, Parker C, Clune J (2018) Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning. Proceedings of the National Academy of Sciences Vol. 115 no. 25. (pdf) (html) (featured on the cover)

2017

• Stefano Nichele chaired the Workshop on Autonomous and Adaptive Systems, 11 May in Oslo including talks on ER.
• Sept 2017: Nicolas Bredèche Tutorial chair at ECAL 2017
• Sept 2017: Proceedings of the Fourteenth European Conference Artificial Life, ECAL 2017 (MIT Press 2017, ISBN 978-0-262-34633-7). Editors: Carole Knibbe, Guillaume Beslon, David P. Parsons, Dusan Misevic, Jonathan Rouzaud-Cornabas, Nicolas Bredèche, Salima Hassas, Olivier Simonin, Hédi Soula.
• Sept 2017: Workshop on evolving physical systems at ECAL 2017. Organizers: John Rieffel, Jean-Baptiste Mouret, Nicolas Bredèche, Evert Haasdijk
• July 2017: Workshop on evolving collective behaviors in robotics at GECCO 2017, Organizers: N. Bredeche, E. Haasdijk, A. Prieto and H. Hamman.
• July 2017: Tutorial on Evolutionary Robotics at GECCO 2017. Organizers: N. Bredeche, S. Doncieux and J-B. Mouret.
• June 2017: special issue on evolving physical systems of ALIFE journal (MIT press). Editors: John Rieffel, Jean-Baptiste Mouret, Nicolas Bredèche, Evert Haasdijk

2016

• Sept.2016: Tutorial chair at PPSN 2016. chairs: N. Bredeche and Carola Doerr

2015

• Stéphane Doncieux initiated a dialogue on “Representational redescription in humans and machines”, in the IEEE CIS Newsletter on CDS, 12(1).
• JB Mouret, J Clune, A. Cully and D. Tarapore published a milestone article in Nature: Cully, A., Clune, J., Tarapore, D., & Mouret, J. B. (2015). Robots that can adapt like animals1. Nature, 521 (7553), 503-507.
• July 2015: Sebastian Risi and Jean-Baptiste Mouret are co-chairing the “Generative and Developmental Systems” track a GECCO’2015 (Madrid)
• April 2015: Sebastian Risi was local chair or Evostar 2015 conference in Copenhagen, Denmark.
• Jean-Baptiste Mouret, Nicolas Bredeche and Stéphane Doncieux presented a 3-hour tutorial about Evolutionary Robotics at GECCO 2015 (Madrid, Spain) and ECAL 2015 (York, UK)
• Sebastian Risi will do a lecture about “Bio-inspired, Automated Design of Machine Bodies and Artificial Brains” at for the Czech European Robotics Week (Prague, November 2015).
• July 2015: Tutorial on Evolutionary Robotics at GECCO 2015 and at ECAL 2015. Organizers: N. Bredeche, S. Doncieux and J-B. Mouret.
• July 2015: Fourth workshop on Evolving Physical Systems at ECAL 2015. Organizers: John Rieffel, Jean-Baptiste Mouret, Nicolas Bredèche, Evert Haasdijk
• July 2015: First workshop on Evolving Collective Behaviours for Robotics at GECCO 2015. Organizers: N. Bredeche, A. Prieto and E. Haasdijk.