REVIEW 12 cited by
Discovering Symbolic Models from Deep Learning with Inductive Biases
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Discovering Symbolic Models from Deep Learning with Inductive Biases
read the original abstract
We develop a general approach to distill symbolic representations of a learned deep model by introducing strong inductive biases. We focus on Graph Neural Networks (GNNs). The technique works as follows: we first encourage sparse latent representations when we train a GNN in a supervised setting, then we apply symbolic regression to components of the learned model to extract explicit physical relations. We find the correct known equations, including force laws and Hamiltonians, can be extracted from the neural network. We then apply our method to a non-trivial cosmology example-a detailed dark matter simulation-and discover a new analytic formula which can predict the concentration of dark matter from the mass distribution of nearby cosmic structures. The symbolic expressions extracted from the GNN using our technique also generalized to out-of-distribution data better than the GNN itself. Our approach offers alternative directions for interpreting neural networks and discovering novel physical principles from the representations they learn.
Forward citations
Cited by 12 Pith papers
-
Sample Complexity of Scientific Discovery: PAC Learnability of Compositional Function Trees
Proves that Rademacher complexity of depth-d compositional trees over finite operator vocabulary is controlled by (K b L)^{d} / sqrt(n) under Lipschitz conditions on operators.
-
Physics-guided discovery of dynamical dark-energy equations of state through iterative AI reasoning
An iterative AI reasoning process proposes new dynamical dark energy equations of state that are competitive with traditional forms on supernova, BAO, and Planck data.
-
EML-CD: Causal Mechanism Recovery via EML Symbolic Trees in Structure Learning
EML-CD recovers causal DAG structure and closed-form mechanisms via gated EML trees, matching PC/GES SHD on Sachs data while recovering 10 of 11 function families in bivariate tests and outperforming SINDy on mechanism f-MSE.
-
Symbolic Classification-Enabled LHC Limits Online BSM Global Fits
Symbolic regression produces an approximate classifier for LHC exclusion limits that enables their direct inclusion during pMSSM global fits.
-
Predicting intermediate-mass black hole formation in star clusters with machine learning
Machine learning regressors trained on Rapster simulations forecast that globular clusters rarely host black holes above 100 solar masses while a few nuclear star clusters may exceed this threshold.
-
Neuro-Symbolic ODE Discovery with Latent Grammar Flow
Latent Grammar Flow discovers ODEs by placing grammar-based equation representations in a discrete latent space, using a behavioral loss to cluster similar equations, and sampling via a discrete flow model guided by d...
-
Learning to Unscramble: Simplifying Symbolic Expressions via Self-Supervised Oracle Trajectories
A permutation-equivariant transformer trained on self-supervised oracle trajectories from scrambled expressions achieves near-perfect simplification rates for dilogarithms and 100% success on 5-point gluon scattering ...
-
Scale-Aware Adversarial Analysis: A Diagnostic for Generative AI in Multiscale Complex Systems
A new scale-aware diagnostic framework shows that unconstrained diffusion generative models exhibit structural freezing and instability instead of smooth physical responses under multiscale perturbations.
-
Into the Gompverse: A robust Gompertzian reionization model for CMB analyses
A Gompertzian reionization model with three nuisance parameters demotes optical depth to a derived quantity, reducing its uncertainty by a factor of three and revealing potential neutrino mass tension in CMB analyses.
-
Self-Revising Discovery Systems for Science: A Categorical Framework for Agentic Artificial Intelligence
A category-theoretic model frames scientific discovery as verified regime transitions via left Kan extensions that preserve and compare artifacts across schema changes in agentic AI.
-
Neuro-Symbolic ODE Discovery with Latent Grammar Flow
Latent Grammar Flow embeds grammar-based ODE representations into a discrete latent space with a behavioural loss and samples candidate equations via discrete flow to fit observed data.
-
Machine Learning for Multi-messenger Probes of New Physics and Cosmology: A Review and Perspective
A review summarizing machine learning methods for multi-messenger probes of dark matter and new physics, with a proposed plan for future integrated analyses.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.