Refine
Document Type
- Article (3)
Language
- English (3)
Has Fulltext
- yes (3) (remove)
Is part of the Bibliography
- no (3) (remove)
Keywords
- active perception (3) (remove)
Institute
The efficient coding hypothesis posits that sensory systems of animals strive to encode sensory signals efficiently by taking into account the redundancies in them. This principle has been very successful in explaining response properties of visual sensory neurons as adaptations to the statistics of natural images. Recently, we have begun to extend the efficient coding hypothesis to active perception through a form of intrinsically motivated learning: a sensory model learns an efficient code for the sensory signals while a reinforcement learner generates movements of the sense organs to improve the encoding of the signals. To this end, it receives an intrinsically generated reinforcement signal indicating how well the sensory model encodes the data. This approach has been tested in the context of binocular vison, leading to the autonomous development of disparity tuning and vergence control. Here we systematically investigate the robustness of the new approach in the context of a binocular vision system implemented on a robot. Robustness is an important aspect that reflects the ability of the system to deal with unmodeled disturbances or events, such as insults to the system that displace the stereo cameras. To demonstrate the robustness of our method and its ability to self-calibrate, we introduce various perturbations and test if and how the system recovers from them. We find that (1) the system can fully recover from a perturbation that can be compensated through the system's motor degrees of freedom, (2) performance degrades gracefully if the system cannot use its motor degrees of freedom to compensate for the perturbation, and (3) recovery from a perturbation is improved if both the sensory encoding and the behavior policy can adapt to the perturbation. Overall, this work demonstrates that our intrinsically motivated learning approach for efficient coding in active perception gives rise to a self-calibrating perceptual system of high robustness.
We present a model for the autonomous and simultaneous learning of active binocular and motion vision. The model is based on the Active Efficient Coding (AEC) framework, a recent generalization of classic efficient coding theories to active perception. The model learns how to efficiently encode the incoming visual signals generated by an object moving in 3-D through sparse coding. Simultaneously, it learns how to produce eye movements that further improve the efficiency of the sensory coding. This learning is driven by an intrinsic motivation to maximize the system's coding efficiency. We test our approach on the humanoid robot iCub using simulations. The model demonstrates self-calibration of accurate object fixation and tracking of moving objects. Our results show that the model keeps improving until it hits physical constraints such as camera or motor resolution, or limits on its internal coding capacity. Furthermore, we show that the emerging sensory tuning properties are in line with results on disparity, motion, and motion-in-depth tuning in the visual cortex of mammals. The model suggests that vergence and tracking eye movements can be viewed as fundamentally having the same objective of maximizing the coding efficiency of the visual system and that they can be learned and calibrated jointly through AEC.
A hypothesis regarding the development of imitation learning is presented that is rooted in intrinsic motivations. It is derived from a recently proposed form of intrinsically motivated learning (IML) for efficient coding in active perception, wherein an agent learns to perform actions with its sense organs to facilitate efficient encoding of the sensory data. To this end, actions of the sense organs that improve the encoding of the sensory data trigger an internally generated reinforcement signal. Here it is argued that the same IML mechanism might also support the development of imitation when general actions beyond those of the sense organs are considered: The learner first observes a tutor performing a behavior and learns a model of the the behavior's sensory consequences. The learner then acts itself and receives an internally generated reinforcement signal reflecting how well the sensory consequences of its own behavior are encoded by the sensory model. Actions that are more similar to those of the tutor will lead to sensory signals that are easier to encode and produce a higher reinforcement signal. Through this, the learner's behavior is progressively tuned to make the sensory consequences of its actions match the learned sensory model. I discuss this mechanism in the context of human language acquisition and bird song learning where similar ideas have been proposed. The suggested mechanism also offers an account for the development of mirror neurons and makes a number of predictions. Overall, it establishes a connection between principles of efficient coding, intrinsic motivations and imitation.