Machine learning for effective quantum-field-theory discovery

Determine whether machine-learning methods can infer effective quantum-field-theory operator bases, couplings, and their scale dependence from lattice configurations, experimental data, or combined datasets while preserving gauge invariance, locality, global symmetries, conservation laws, and the correct continuum limit.

Background

The paper distinguishes theory discovery from conventional parameter estimation. In this proposed direction, machine learning would infer the effective theory itself—including relevant operators, coupling constants, and renormalization-group evolution—instead of merely reconstructing observables or accelerating simulations for a prescribed action.

The problem remains difficult because different operator bases may describe the same finite-resolution data, while learned Hamiltonians can suffer from non-identifiability, finite-volume effects, discretization artifacts, and statistical uncertainty. A successful solution would need independent-observable tests, stability across resolution scales, symmetry checks, and consistency with known continuum limits.

References

A more ambitious direction is to reverse this logic: can machine learning infer the effective theory itself from physical data?

— Machine Learning Meets High-Energy Nuclear Physics: From Pattern Recognition to Physics-Integrated Discovery  (2610.12293 - Chen et al., 8 Oct 2026) in Section 9, 'Learning Quantum Field Theories Across Scales'

Can AI connect sectors that were historically analyzed separately? This is the central question for unified QCD inference.

— Machine Learning Meets High-Energy Nuclear Physics: From Pattern Recognition to Physics-Integrated Discovery  (2610.12293 - Chen et al., 8 Oct 2026) in Section 11, Conclusions and outlook