Robot self/other distinction: active inference meets neural networks learning in a mirror

@inproceedings{Lanillos2020RobotSD,
  title={Robot self/other distinction: active inference meets neural networks learning in a mirror},
  author={Pablo Lanillos and Jordi Pag{\`e}s and Gordon Cheng},
  booktitle={ECAI},
  year={2020}
}
Self/other distinction and self-recognition are important skills for interacting with the world, as it allows humans to differentiate own actions from others and be self-aware. However, only a selected group of animals, mainly high order mammals such as humans, has passed the mirror test, a behavioural experiment proposed to assess self-recognition abilities. In this paper, we describe self-recognition as a process that is built on top of body perception unconscious mechanisms. We present an… 

Figures from this paper

Robot in the Mirror: Toward an Embodied Computational Model of Mirror Self-Recognition
TLDR
The core of the technical contribution is learning the appearance representation and visual novelty detection by means of learning the generative model of the face with deep auto-encoders and exploiting the prediction error.
Enabling the Sense of Self in a Dual-Arm Robot
TLDR
A neural network architecture is implemented to enable a dual-arm robot to differentiate its limbs from the environment using visual and proprioception sensory inputs and demonstrates experimentally that a robot can distinguish itself with an accuracy of 88.7% on average in cluttered environmental settings and under confounding input signals.
Sensorimotor representation learning for an "active self" in robots: A model survey
TLDR
A theoretical computational framework is proposed, which aims to allow the emergence of the sense of self in artificial agents by developing sensory representations through self-exploration.
Neuroscience-inspired perception-action in robotics: applying active inference for state estimation, control and self-perception
TLDR
How neuroscience findings open up opportunities to improve current estimation and control algorithms in robotics is discussed, and active inference, a mathematical formulation of how the brain resists a natural tendency to disorder, provides a unified recipe to potentially solve some of the major challenges in robotics.
The predictive brain in action: Involuntary actions reduce body prediction errors
TLDR
The present study shows that humans move their body to match the expected location according to other sensory input, which corresponds to reducing the prediction error, which means that the brain adapts to conflicting or uncertain information from the senses by unconsciously acting in the world.
Modeling robot co-representation: state-of-the-art, open issues, and predictive learning as a possible framework
  • M. Kirtay, Olga A. Wudarczyk, V. Hafner
  • Biology, Psychology
    2020 Joint IEEE 10th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)
  • 2020
TLDR
An integrated view of the state-of-the-art findings in robotics literature on co-representation is offered and directions for future research are outlined, with the aim to boost success in building robots equipped with co- Representation models fit for smooth social interactions.
Predictive Processing in Cognitive Robotics: A Review
TLDR
Analysis of cognitive robotics research working under the predictive processing framework suggests that research in both cognitive robotics implementations and nonrobotic models needs to be extended to the study of how multiple exteroceptive modalities can be integrated into prediction error minimization schemes.
Active strategies for multisensory conflict suppression in the virtual hand illusion
TLDR
A computational model, based on the free-energy principle, that forecasts involuntary movements in sensorimotor conflicts is described and experimentally confirmed in humans using a virtual reality rubber-hand illusion, indicating that the brain-body is generating actions to reduce the prediction error between the expected arm location and the new visual arm.
Robotic self-representation improves manipulation skills and transfer learning
TLDR
This paper develops a model that learns bidirectional action-effect associations to encode the representations of body schema and the peripersonal space from multisensory information, which is named multimodal BidAL and demonstrates that it significantly stabilizes the learning-based problem-solving under noisy conditions and improves transfer learning of robotic manipulation skills.
...
...

References

SHOWING 1-10 OF 39 REFERENCES
A Predictive Processing Model of Perception and Action for Self-Other Distinction
TLDR
Results from simulations of different scenarios show that the proposed hierarchical model of sensorimotor coordination is not only able to minimize free energy during perception and action, but also showing that the model can correctly attribute sense of agency to own actions.
Emergence of mirror neuron system: Immature vision leads to self-other correspondence
TLDR
A computational model is proposed for early development of the mirror neuron system, which enables a robot to imitate others' actions, which is originated in immature vision.
Active inference body perception and action for humanoid robots
TLDR
This work presents a humanoid robot interacting with the world by means of a human brain-like inspired perception and control algorithm based on the free-energy principle and shows the advantages of such an algorithm for real world applications.
Yielding Self-Perception in Robots Through Sensorimotor Contingencies
TLDR
A self-perception mechanism that enables a humanoid robot to understand certain sensory changes caused by naive actions during interaction with objects is presented and a novel model for self-detection, which does not need to track the body parts is presented.
Using probabilistic reasoning over time to self-recognize
From the perception of action to the understanding of intention
TLDR
Evidence that biological motion is processed as a special category, to which humans from an early age attribute mental states such as intention is reviewed, to support the idea that the brain is a powerful simulating machine, designed to extract intentions from the motion and to predict the future actions of other animate beings.
Adaptive Robot Body Learning and Estimation Through Predictive Coding
  • Pablo Lanillos, G. Cheng
  • Computer Science
    2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
  • 2018
TLDR
A computational perceptual model based on predictive processing that enables any multisensory robot to learn, infer and update its body configuration when using arbitrary sensors with Gaussian additive noise is introduced.
Self-detection in robots: a method based on detecting temporal contingencies†
TLDR
A methodology for autonomous learning of the characteristic delay between motor commands and observed movements of visual stimuli and it is described that the robot estimates its own efferent-afferent delay from self-observation data gathered while performing motor babbling.
A match does not make a sense: on the sufficiency of the comparator model for explaining the sense of agency
TLDR
An invalidating theoretical flaw is identified in the reasoning underlying this conversion from adult studies to developmental science and cognitive systems research, rooted in an oversight in the conceptualization of the comparator model as currently used in experimental practice.
Action and behavior: a free-energy formulation
TLDR
The free-energy formulation may provide an alternative perspective on the motor control that places it in an intimate relationship with perception, and can explain why adaptive behavior emerges in biological agents and suggests a simple alternative to optimal control theory.
...
...