Abstract
Curiosity Driven Modular Incremental Slow Feature Analysis (CD-MISFA;) is a recently introduced model of intrinsically-motivated invariance learning. Artificial curiosity enables the orderly formation of multiple stable sensory representations to simplify the agent's complex sensory input. We discuss computational properties of the CD-MISFA model itself as well as neurophysiological analogs fulfilling similar functional roles. CD-MISFA combines 1. unsupervised representation learning through the slowness principle, 2. generation of an intrinsic reward signal through learning progress of the developing features, and 3. balancing of exploration and exploitation to maximize learning progress and quickly learn multiple feature sets for perceptual simplification. Experimental results on synthetic observations and on the iCub robot show that the intrinsic value system is essential for representation learning. Representations are typically explored and learned in order from least to most costly, as predicted by the theory of curiosity.
Highlights
We describe a model called CURIOUSity-DRiven, Modular, Incremental Slow Feature Analysis (Curious Dr MISFA), which autonomously explores various action contexts, learning low-dimensional encodings from the high-dimensional sensory inputs that result from each such context
We show why the interacting subsystems of CD-MISFA are necessary for the kind of unsupervised learning it undertakes, but we show how the subsystems that enable the model to autonomously explore and acquire new sensory representations, mirror the functional roles of some of the underlying cortical and neuromodulatory systems responsible for unsupervised learning, intrinsic motivation, task engagement, and task switching
In the absence of external rewards, how should an agent decide which actions and contexts to explore, in order to determine which representations are relevant and learnable? If a sensory representation is deemed overly complex or even unlearnable, what are the mechanisms by which the agent can disengage from exploring its current context, in order to allow it to explore others? CD-MISFA is an algorithmic approach to developmental robotics, and does not explicitly model the neural mechanisms by which these functions are realized in the brain, it is notable that the functional roles of the various subsystems in CD-MISFA find counterparts in neurophysiology
Summary
We describe a model called CURIOUSity-DRiven, Modular, Incremental Slow Feature Analysis (Curious Dr MISFA), which autonomously explores various action contexts, learning low-dimensional encodings from the high-dimensional sensory inputs (i.e., video) that result from each such context. Autonomous behavior in this regard requires the coordinated interaction between a number of subsystems which enable an agent to balance exploration-exploitation, to engage in useful contexts while disengaging from others, and to organize representations such that newly learned representations do not overwrite previously learned ones. Curious Dr MISFA representations, again in the predicted order, all while operating on high-dimensional video streams as sensory input
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
