AI Feynman: A physics-inspired method for symbolic regression

@article{Udrescu2020AIFA,
  title={AI Feynman: A physics-inspired method for symbolic regression},
  author={Silviu-Marian Udrescu and Max Tegmark},
  journal={Science Advances},
  year={2020},
  volume={6}
}
Our physics-inspired algorithm for symbolic regression is able to discover complex physics equations from mere tables of numbers. A core challenge for both physics and artificial intelligence (AI) is symbolic regression: finding a symbolic expression that matches data from an unknown function. Although this problem is likely to be NP-hard in principle, functions of practical interest often exhibit symmetries, separability, compositionality, and other simplifying properties. In this spirit, we… 
AI Feynman 2.0: Pareto-optimal symbolic regression exploiting graph modularity
We present an improved method for symbolic regression that seeks to fit data to formulas that are Pareto-optimal, in the sense of having the best accuracy for a given complexity. It improves on the
Neural Symbolic Regression that Scales
TLDR
This paper procedurally generate an unbounded set of equations, and simultaneously pre-train a Transformer to predict the symbolic equation from a corresponding set of input-output-pairs, and shows empirically that this approach can re-discover a set of well-known physical equations and that it improves over time with more data and compute.
Symbolic Regression for Space Applications: Differentiable Cartesian Genetic Programming Powered by Multi-objective Memetic Algorithms
TLDR
A new multi-objective memetic algorithm is proposed that ex-ploits a di-erentiable Cartesian Genetic Programming encoding to learn constants during evolutionary loops and shows that this approach is competitive or outperforms machine learned black box regression models or hand-engineered models for two applications from space.
A Neuro-Symbolic Method for Solving Differential and Functional Equations
TLDR
This work introduces a method for generating symbolic expressions to solve differential equations while leveraging deep learning training methods, and proposes a novel neural architecture for learning mathematical expressions to optimize a customizable objective.
Deep symbolic regression: Recovering mathematical expressions from data via policy gradients
TLDR
This work proposes a framework that combines deep learning with symbolic regression via a simple idea: use a large model to search the space of small models, and employ reinforcement learning to train the network to generate better-fitting expressions.
Automated Symbolic Law Discovery: A Computer Vision Approach
TLDR
Inspired by the incredible success of deep learning in computer vision, this work tackles the problem of automatically discovering scientific laws from experimental data by adapting various successful network architectures into the symbolic law discovery pipeline, and significantly outperforms the current state of the art in terms of data fitting, discovery rate, and succinctness.
Fast Neural Models for Symbolic Regression at Scale
TLDR
This work introduces OccamNet, a neural network model that finds interpretable, compact, and sparse solutions for fitting data, à la Occam’s razor, and introduces a two-step optimization method that samples functions and updates the weights with backpropagation based on cross-entropy matching in an evolutionary strategy.
End-to-end symbolic regression with transformers
TLDR
This paper challenges this two-step procedure, and task a Transformer to directly predict the full mathematical expression, constants included, and presents ablations to show that this end-to-end approach yields better results, sometimes even without the refinement step.
CAL EXPRESSIONS FROM DATA VIA POLICY GRADIENTS
TLDR
This work proposes a framework that combines deep learning with symbolic regression via a simple idea: use a large model to search the space of small models, and employ reinforcement learning to train the network to generate better-fitting expressions.
AI Descartes: Combining Data and Theory for Derivable Scientific Discovery
TLDR
A method for combining logical reasoning with symbolic regression is developed, enabling principled derivations of models of natural phenomena and envision that this combination will enable derivable discovery of fundamental laws of science.
...
...

References

SHOWING 1-10 OF 56 REFERENCES
FFX: Fast, Scalable, Deterministic Symbolic Regression Technology
TLDR
A new non-evolutionary technique for symbolic regression that is orders of magnitude faster than competent GP approaches on real-world problems, returns simpler models, has comparable or better prediction on unseen data, and converges reliably and deterministically.
Distilling Free-Form Natural Laws from Experimental Data
TLDR
This work proposes a principle for the identification of nontriviality, and demonstrated this approach by automatically searching motion-tracking data captured from various physical systems, ranging from simple harmonic oscillators to chaotic double-pendula, and discovered Hamiltonians, Lagrangians, and other laws of geometric and momentum conservation.
Why Does Deep and Cheap Learning Work So Well?
TLDR
It is argued that when the statistical process generating the data is of a certain hierarchical form prevalent in physics and machine learning, a deep neural network can be more efficient than a shallow one.
Discovering governing equations from data by sparse identification of nonlinear dynamical systems
TLDR
This work develops a novel framework to discover governing equations underlying a dynamical system simply from data measurements, leveraging advances in sparsity techniques and machine learning and using sparse regression to determine the fewest terms in the dynamic governing equations required to accurately represent the data.
A Statistical Learning Perspective of Genetic Programming
TLDR
It is proved that a parsimonious fitness ensures universal consistency and a more complicated modification of the fitness is proposed in order to avoid unnecessary bloat while nevertheless preserving universal consistency.
Automated refinement and inference of analytical models for metabolic networks.
TLDR
An automated method is demonstrated to simplify the reverse engineering of metabolic networks from experimental data by modifying an existing or related model by modifying nonlinear terms and structural modifications or even constructing a new model that agrees with the system's time series observations.
Genetic programming - on the programming of computers by means of natural selection
  • J. Koza
  • Computer Science
    Complex adaptive systems
  • 1993
TLDR
This book discusses the evolution of architecture, primitive functions, terminals, sufficiency, and closure, and the role of representation and the lens effect in genetic programming.
Sparse identification of nonlinear dynamics for rapid model recovery.
TLDR
The proposed abrupt-SINDy architecture provides a new paradigm for the rapid and efficient recovery of a system model after abrupt changes, and shows that sparse updates to a previously identified model perform better with less data, have lower runtime complexity, and are less sensitive to noise than identifying an entirely new model.
Genetic Algorithms for Pattern Recognition
TLDR
Genetic Algorithms for Pattern Recognition covers a broad range of applications in science and technology, describing the integration of genetic algorithms in pattern recognition and machine learning problems to build intelligent recognition systems.
Genetic Algorithms and Investment Strategies
TLDR
Recognizing the continued resistance of many traders and analysts to GAs, Richard Bauer's Genetic Algorithms and Investment Strategies shows how these approaches do not herald an age in which people will be supplanted by machines, revealing instead how they serve only to augment human thinking.
...
...