Research programs and open obstacles

The main research programs attack different parts of the Yang–Mills problem. Some control ultraviolet fluctuations; others constrain lattice observables, explore infrared mechanisms, or extract numerical spectra. Their results become comparable when we ask what they establish, in which theory, and through which limits.

This page introduces the enduring strategies. The Research Atlas holds the dated results and their current source records.

Prerequisites: the course’s research-reading lesson or equivalent experience with regulated field theory. Outcome: choose a method by the quantity it controls, then locate its precise remaining implication.

Work through a method	Main question
Constructive renormalization	Which scale-dependent estimates survive many integrations?
Rigorous lattice methods	Which hypotheses imply a particular correlation or loop bound?
Stochastic quantization	How do local dynamics, invariant measures and physical reconstruction differ?
Lattice bootstrap	What can exact finite constraints and certified optimization establish?
Semiclassical methods	Which hierarchy makes a mechanism calculable, and what continuation remains?

Each guide includes a worked example and a solved exercise. The open-obstacle index connects their questions to the same target without ranking approaches.

Constructive renormalization

Constructive RG studies how an effective description changes as fluctuations at successive scales are integrated out. Its estimates must control the field configurations that matter, including large-field regions, while tracking the dependence on the regulator.

Bałaban’s conditional ultraviolet-stability result requires all effective couplings to stay sufficiently small throughout the stated construction. Its ultraviolet control does not establish that premise along the desired trajectory, a completed continuum correlation theory, or a physical gap. The outstanding bridge includes identifying the limit and controlling its infrared behavior.

Prepare with distributions and uniform estimates.

Stochastic approaches

Stochastic quantization introduces an auxiliary evolution designed to sample a field measure. The analysis can establish existence of dynamics, renormalization rules, invariant measures, or correlation estimates.

These conclusions are distinct. A local stochastic solution need not already yield a global invariant measure; a relaxation gap is not automatically the physical Hamiltonian gap. The semigroup bridge explains why.

Chevyrev and Shen’s three-dimensional renormalization result supplies a precise recent example. Its dimension, matter content, and local-solution setting are part of its meaning.

Lattice inequalities and confinement criteria

Correlation methods can turn controlled estimates into theorems about Wilson loops and confinement in specified lattice regimes. Chatterjee’s criterion and the 2025 dynamical area-law result make this strategy concrete.

The next question is whether the required assumptions can be established along the continuum scaling trajectory. A strong-coupling theorem remains useful even when that extension is open.

Positivity bootstrap and loop equations

The starting point is elementary:

$$\left\langle \left(\sum_i c_iO_i\right)^\dagger \left(\sum_j c_jO_j\right) \right\rangle\geq0.$$

Thus the matrix $M_{ij}=\langle O_i^\dagger O_j\rangle$ is positive semidefinite. Combine such constraints with loop equations and trace identities to bound selected observables.

The SU(3) dossier and direct and indirect loop equations show current applications. Important obstacles include reaching sufficiently long distances, understanding truncation, certifying numerical optimization, and controlling the continuum limit.

Spectroscopy and controlled mechanisms

Lattice spectroscopy supplies quantitative evidence for the physical spectrum through simulations and continuum extrapolations. Its statistical and systematic uncertainties belong in the result.

Semiclassical compactifications instead expose analytic mechanisms in specified regimes. Their transfer to undeformed four-dimensional space requires its own argument.

Functional equations, Hamiltonian methods, generalized symmetries, large $N$, and holographic descriptions add further perspectives. The useful question is always what observable or obstruction they control and which assumptions survive the intended limit.

Build your research route

Choose a question in the Proof Map, read the corresponding introductory lessons, and then study one result dossier. Write down its strongest conclusion and a single concrete missing step. Reading a research paper gives a reusable procedure.

The cited dossiers link directly to primary sources. Comparing methods through their actual hypotheses preserves both the significance of partial progress and the difficulty of the remaining problem.