The Physics of Learning
arXiv:2209.11954 · doi:10.1080/00107514.2022.2135672
Abstract
A learning machine, like all machines, is an open system driven far from thermal equilibrium by access to a low entropy source of free energy. We discuss the connection between machines that learn, with low probability of error, and the optimal use of thermodynamic resources for both classical and quantum machines. Both fixed point and spiking perceptrons are discussed in the context of possible physical implementations. An example of a single photon quantum kernel evaluation illustrates the important role for quantum coherence in data representation. Machine learning algorithms, implemented on conventional complementary metal oxide semiconductor (CMOS) devices, currently consume large amounts of energy. By focusing on the physical constraints of learning machines rather than algorithms, we suggest that a more efficient means of implementing learning may be possible based on quantum switches operating at very low power. Single photon kernel evaluation is an example of the energy efficiency that might be possible.
arXiv comment44 pages, 27 figures
References in corpus (22)
- A nonequilibrium equality for free energy differences
- Stochastic thermodynamics, fluctuation theorems, and molecular machines
- Machine learning and the physical sciences
- Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systems
- Quantum advantage in learning from experiments
- An optical neural network using less than 1 photon per multiplication
- Heralded Storage of a Photonic Quantum Bit in a Single Atom
- Temporal Modes in Quantum Optics: Then and Now
- Atomic vapor-based high efficiency optical detectors with photon number resolution
- Sensitive radio-frequency measurements of a quantum dot by tuning to perfect impedance matching
- Introducing one-shot work into fluctuation relations
- Autonomous tuning and charge state detection of gate defined quantum dots
- The thermodynamics of clocks
- Detailed Fluctuation Theorems: A Unifying Perspective
- Quantum Noise in the Electromechanical Shuttle
- TherML: Thermodynamics of Machine Learning
- Quantum Mean Embedding of Probability Distributions
- A Physical Quantum Agent
- Thermodynamic efficiency of learning a rule in neural networks
- Classical and Quantum Causal Interventions
- Asymptotic work distributions in driven bistable systems
- Conditional counting statistics of electrons tunneling through quantum dot systems measured by a quantum point contact