ICDL 2022 will feature an exciting list of keynote speakers (3/3):
Ryota Kanai @kanair (Araya, Japan)
José Santos-Victor @Jose_SV_ (Instituto Superior Técnico, Portugal)
Aikaterini Fotopoulou (University College London, UK)
ICDL 2022 will feature an exciting list of keynote speakers (2/3):
Lars Chittka @LChittka (Queen Mary University of London, UK)
Barbara Webb (University of Edinburgh, UK)
ICDL 2022 will feature an exciting list of keynote speakers (1/3):
Anil Seth @anilkseth (University of Sussex, UK)
Raia Hadsell @RaiaHadsell (Google DeepMind, UK)
You can follow most of the talks of @ICDL_EpiRob 2020 on YouTube under the playlist below. We will select the best presentation award based on the likes of the videos.
https://t.co/NzAhNXpzch
@DevRobNews
We just posted a clarification regarding the schedule of @ICDL_EpiRob 2020. The conference is scheduled for 26-30 October, and it starts every day at 7:00 AM UTC-4. The iCalendar would automatically show the time in your local time zone.
https://t.co/i35yBPIHhk
@DevRobNews
Kevin O'Regan highlighting problems in identifying learning of sensorimotor contingencies in > 300 infants. Pointing out extensive follow up work on identifying shortcoming of experimental design and asking #imol19 attendees for further ideas. Great transparency, great science!
Jochen Triesch using efficient coding of stimuli in agent representations as intrinsic reward to drive behavior, and e.g. model dev. of binocular vision. Wondering if this could yield representations of time & space as in futurist/cubist painting. #imol19#computationalcreativity
Day 2 at #imol19. @YukieNagai investigating #predictivecoding for the cognitive development of infants: #activeinference giving rise to proto-social behaviour; properties of priors could explain non-neurotypical behaviours. Really smart experiments!
Bruno Castro da Silva presenting two approaches to identify reusable behaviours as #RL options: 'policy compression' and 'early life options'. The latter selected to perform well on random policy of early learning agent. Outperforming atomic action #RL up to 200k steps. #imol19
At the 4th Int. Workshop on Intrinsically Motivated Open-ended Learning to bridge between communities. Keen to learn about their #computationalcreativity#gameai heritage and ideas, and discuss key issues on defining, evaluating and expanding #intrinsicmotivation. #imol19
Please share:
Special Issue on Continual Unsupervised Sensorimotor Learning
Submission deadline: 15th March 2019
Expected publication: September 2019
https://t.co/sJX0x7kFDv