Zobrazeno 1 - 10
of 45
pro vyhledávání: '"Coninx, Alexandre"'
Publikováno v:
GECCO 22 Companion, July 9-13, 2022, Boston, MA, USA
The Novelty Search (NS) algorithm was proposed more than a decade ago. However, the mechanisms behind its empirical success are still not well formalized/understood. This short note focuses on the effects of the archive on exploration. Experimental e
Externí odkaz:
http://arxiv.org/abs/2205.03162
Publikováno v:
IEEE Robotics and Automation Letters 7.2 (2022): 4424-4431
In the past few years, a considerable amount of research has been dedicated to the exploitation of previous learning experiences and the design of Few-shot and Meta Learning approaches, in problem domains ranging from Computer Vision to Reinforcement
Externí odkaz:
http://arxiv.org/abs/2109.06826
As open-ended learning based on divergent search algorithms such as Novelty Search (NS) draws more and more attention from the research community, it is natural to expect that its application to increasingly complex real-world problems will require t
Externí odkaz:
http://arxiv.org/abs/2104.03936
Reward-based optimization algorithms require both exploration, to find rewards, and exploitation, to maximize performance. The need for efficient exploration is even more significant in sparse reward settings, in which performance feedback is given s
Externí odkaz:
http://arxiv.org/abs/2102.03140
Autor:
Doncieux, Stephane, Bredeche, Nicolas, Goff, Léni Le, Girard, Benoît, Coninx, Alexandre, Sigaud, Olivier, Khamassi, Mehdi, Díaz-Rodríguez, Natalia, Filliat, David, Hospedales, Timothy, Eiben, A., Duro, Richard
Robots are still limited to controlled conditions, that the robot designer knows with enough details to endow the robot with the appropriate models or behaviors. Learning algorithms add some flexibility with the ability to discover the appropriate be
Externí odkaz:
http://arxiv.org/abs/2005.06223
Evolvability is an important feature that impacts the ability of evolutionary processes to find interesting novel solutions and to deal with changing conditions of the problem to solve. The estimation of evolvability is not straightforward and is gen
Externí odkaz:
http://arxiv.org/abs/2005.06224
Autor:
Merckling, Astrid, Coninx, Alexandre, Cressot, Loic, Doncieux, Stéphane, Perrin-Gilbert, Nicolas
Robots could learn their own state and world representation from perception and experience without supervision. This desirable goal is the main focus of our field of interest, state representation learning (SRL). Indeed, a compact representation of s
Externí odkaz:
http://arxiv.org/abs/1910.01738
Performing Reinforcement Learning in sparse rewards settings, with very little prior knowledge, is a challenging problem since there is no signal to properly guide the learning process. In such situations, a good search strategy is fundamental. At th
Externí odkaz:
http://arxiv.org/abs/1909.05508
Robots need to understand their environment to perform their task. If it is possible to pre-program a visual scene analysis process in closed environments, robots operating in an open environment would benefit from the ability to learn it through the
Externí odkaz:
http://arxiv.org/abs/1903.04413
To solve its task, a robot needs to have the ability to interpret its perceptions. In vision, this interpretation is particularly difficult and relies on the understanding of the structure of the scene, at least to the extent of its task and sensorim
Externí odkaz:
http://arxiv.org/abs/1901.10968