R-IAC: Robust Intrinsically Motivated Exploration and Active Learning

By Adrien Baranes, Pierre-Yves Oudeyer. 2009

150 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Control and Systems Engineering

Citations in the corpus

Towards Continual Reinforcement Learning: A Review and Perspectives (2020, methodology): “Meta-learning has been repeatedly used in the recent literature to learn functions for intrinsic motivation and improved exploration (Baranes and Oudeyer, 2009; Zheng, Oh, and Singh, 2018; Stadie, Yang, Houthooft, Chen, Duan, Wu, Abbeel, and Sutskever, 2018; Xu, Liu, Zhao, and Peng, 2018b; Houthooft”
GRAIL: A Goal-Discovering Robotic Architecture for Intrinsically-Motivated Learning (2016, background): “that tries to anticipate the effects of the actions of the artificial system in the environment [10], [31], [43], [57], [60].”
On-line regression algorithms for learning mechanical models of robots: A survey (2011, background): “So far, there are few attempts to design speciﬁc active learning meth-ods for learning mechanical models of robots, exceptions being Baranes & Oudeyer [5, 6], Martinez-Cantin et al. [64], Robbel [82], Saegusa et al. [85].”
Programmation et apprentissage bayésien pour les jeux vidéo multi-joueurs, application à l'intelligence artificielle de jeux de stratégies temps-réel. (Bayesian Programming and Learning for Multi-Player Video Games, Application to RTS AI) (2012, background): “Another interesting perspective would be to use a hierarchy (from strategy to tactics) of reinforcement learning of (all) the bot’s parameters, which can be seen as learning the degrees of liberty (of the whole bot) one by one as in [Baranès and Oudeyer, 2009].”
Intrinsically motivated goal exploration for active motor learning in robots: A case study (2010, methodology): “Actually, they do not consider the notion of goal, or task, and only try to improve the quality of forward models with no consideration about how they can be reused for control (this applies for instance to IAC [1] and RIAC [ 2 ]).”
Towards Deep Developmental Learning (2016, background): “In [151], this clustering is performed hierarchically.”
Active Learning and Intrinsically Motivated Exploration in Robots : Advances and Challenges (2010, background): “As argued in [19], [24], and [38], architectures based on intrinsically motivated learning can be conceptualized as active learning mechanisms which, in addition to allowing for the self-organized formation of behavioral and developmental complexity, can also allow an agent to efficient learn a…”
The strategic student approach for life-long exploration and learning (2012, methodology): “As a generalization of exploration methods in re-inforcement learning, such as [21], ideas have been suggested such as planning to be surprised [59] or the combination of empirical learning progress with visit counts [60].”
A comprehensive survey on safe reinforcement learning (2015, background): “The exploration bonuses adds a bonus to states with higher potential of learning (Baranes and Oudeyer, 2009), or with higher uncertainty (Brafman and Tennenholtz, 2003).”
Learning and Using Models (2012, methodology): “Baranes and Oudeyer (2009) present an algorithm based on a similar idea called Robust Intelligent Adaptive Curiosity (R-IAC), a method for providing intrinsic reward to encourage a developing agent to explore.”
A free energy principle for a particular physics (2019, background): “In cognitive neuroscience, this ultimately leads to ambiguity reducing, curious behaviour that underwrites the exploratory (i.e., epistemic) foraging that characterises self-evidencing (Baranes and Oudeyer, 2009; Berlyne, 1950; Friston et al., 2017b; Schmidhuber, 2006; Still and Precup, 2012).”
Active Inference, Curiosity and Insight (2017, background): “In neurorobotics, these imperatives are often addressed in terms of active learning (Markant & Gureckis, 2014; Markant, Settles, & Gureckis, 2016), with a focus on intrinsic motivation (Baranes & Oudeyer, 2009).”
The interaction of maturational constraints and intrinsic motivations in active motor development (2011, methodology): “close to 0) represents a system that is competent in reaching the goal � while respecting requirements p. A typical instantiation of ::Jim, without requirements, is defined as 8im( s�, sj, 0) = -II s� -sj 11 2, and is the direct transposition of prediction error in RIAC (which here refers to goal re”
Guest Editorial Active Learning and Intrinsically Motivated Exploration in Robots: Advances and Challenges (2010, background): “As argued in [19], [24], and [38], architectures based on intrinsically motivated learning can be conceptualized as active learning mechanisms which, in addition to allowing for the self-organized formation of behavioral and developmental complexity, can also allow an agent to efﬁcient learn a…”
Intrinsically Motivated Learning of Real-World Sensorimotor Skills with Developmental Constraints (2013, methodology): “ACTUATOR-RIAC corresponds to the original version of R- IAC (Baranes and Oudeyer, 2009) that actively generates actions inside the same space of synergies as ACTUATOR-RANDOM.”