framework
active
framework:physical-learning

Physical learning

Framework for solving inverse problems in which physical systems autonomously adapt their parameters in response to stimuli through local learning rules, without requiring computational design or explicit cost functions

Neighborhood — ranked by edge-count

Papers (1)

paper

Methods (1)

method
  • Local learning rule
    implements
    Learning rule where change in a parameter at point x,t depends only on system state at same or nearby spacetime points, without requiring global cost function computation

Concepts (4)

concept
  • Elastic networks
    associated_with
    Spring networks and mechanical systems used to demonstrate physical learning through modifications of bond stiffnesses in response to strain
  • Flow networks
    associated_with
    Networks for transport via fluid flow (biological vascular, engineered microfluidic); pipe properties adapt locally to control flow conductance and enable learning
  • Origami and Kirigami sheets
    associated_with
    2D sheets with crease patterns that fold into specific topologies; learning d.o.f include crease bending stiffness; can be trained for classification tasks
  • Inverse Problem
    implements
    Problem of finding a system with desired response to perturbation; contrasted with forward problem of predicting responses of a given system

Claims (1)

claim

Frameworks (2)

framework
  • Contrastive learning
    associated_with
    Supervised learning framework where system learns by observing contrast between current response and nudged improved response; requires weak additional forces from supervisor
  • Neuromorphic computing
    associated_with
    Related field where physical elements are modified for desired computational ability; traditionally targets symbolic inputs/outputs unlike physical learning's physical stimuli/responses

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

  • Learningconcept0.840
    Inference of parameters encoding contingencies of the world (e.g., likelihood matrix A) at slower timescale than perception.
  • Perceptual Learningconcept0.826
    Process of inferring causes of sensory information; unified with value learning as integral aspects of free energy minimization.
  • Q-learningmethod0.795
    Model-free RL algorithm used in experimental comparison; employs ε-greedy exploration.
  • Concept Learningconcept0.793
    Acquisition of new concepts by Bayesian model expansion and reduction.
  • Early learningconcept0.787
    The primary domain in which Nicholson's Theory of Loose Parts became influential and known.
  • Associative Learningconcept0.781
    Sentience criterion; capacity occurs even in gene regulatory networks and non-neural morphogenetic agents.
  • Physical instanceconcept0.775
    The view that the individual is a particular piece of hardware running the model over a given period of time
  • Multi-Task Learningconcept0.771
    Learning paradigm that jointly learns multiple related tasks using a single model