# Lax–Oleinik Aubry contact probe

- ID: 3014
- Canonical URL: https://synthcore.org/idea/3014/lax-oleinik-aubry-contact-probe
- API JSON: https://synthcore.org/api/idea/3014.json
- API Markdown: https://synthcore.org/api/idea/3014.md
- Verification status: unverified
- Source: [arXiv:2609.01557](https://arxiv.org/abs/2609.01557)
- Category: geometry
- Solves: accuracy, sample-efficiency, generalization
- ML areas: world-model, rl, training, data-augmentation
- Math tags: dynamical-systems, geometry, pde, convex-analysis, optimization
- Ratings: usefulness 6/10; difficulty 6/10; novelty 8/10

## Idea description

Use the discrepancy between a learned potential and its long-time backward Lax–Oleinik evolution to identify dynamically critical states. Persistent near-contact points are candidates for the Aubry set and can guide adaptive collocation, while states with large gaps can receive fewer training samples.

## Mathematical statement

The paper characterizes the projected Aubry set through the limiting potential u_\infty and its associated backward weak KAM solution u_-: \mathcal{A}=\{x:u_\infty(x)=u_-(x)\}, equivalently \mathcal{A}=\{x:Du_\infty(x)    ext{ exists and }H(Du_\infty(x),x)=c\}. For a Tonelli Hamiltonian H, its Legendre transform is L(v,x)=\sup_{p\in\mathbb{R}^n}\{p\cdot v-H(p,x)\}, where v is velocity and p is momentum. The backward Lax–Oleinik operator is (T_t^-f)(x)=\inf_{\gamma(t)=x}[f(\gamma(0))+\int_0^tL(\dot\gamma(s),\gamma(s))ds]. A backward weak KAM solution is invariant under this evolution up to the critical normalization, T_t^-u_-=u_-+ct. The neural diagnostic is therefore g_t(x)=u_    heta(x)-(T_t^-u_    heta)(x)+ct; small persistent absolute gap indicates a candidate contact or Aubry point.

## Key formulas

- $$\mathcal{A}=\{x\in\mathbb{T}^{n}:u_{\infty}(x)=u_{-}(x)\}=\{x\in\mathbb{T}^{n}:Du_{\infty}(x)\ \text{exists and }H(Du_{\infty}(x),x)=c\}.$$
- $$L(v,x)=\sup_{p\in\mathbb{R}^{n}}\{p\cdot v-H(p,x)\}.$$
- $$(T_t^-f)(x)=\inf_{\substack{\gamma:[0,t]\to\mathbb{T}^n\\\gamma(t)=x}}\left[f(\gamma(0))+\int_0^tL(\dot\gamma(s),\gamma(s))\,ds\right].$$
- $$g_t(x)=u_\theta(x)-(T_t^-u_\theta)(x)+ct,\qquad \mathcal{A}_{\mathrm{probe}}(\varepsilon)=\{x:|g_t(x)|\le\varepsilon\}.$$

## Implementation notes

Attach this module between neural PDE training iterations, after the network has produced a preliminary value function. Evaluate u_theta on a dense periodic candidate grid and approximate T_t^-u_theta at each endpoint. For a first MVP, discretize time into M steps of size Delta t and restrict velocities to a finite stencil V. Initialize A_0(y)=u_theta(y), then apply the dynamic-programming update A_{r+1}(x)=min_{v in V}[A_r(x-Delta t v)+Delta t L(v,x-Delta t v)], using periodic indexing. Set A_M(x) as the approximation to T_t^-u_theta(x), compute g_t(x)=u_theta(x)-A_M(x)+ct, and subtract its median to remove numerical additive drift. Define contact weights q_i=exp(-|g_t(x_i)|/tau). Use a 50/50 mixture of uniform collocation points and points sampled proportional to q_i; alternatively, add an auxiliary penalty mean_i min(|g_t(x_i)|,epsilon)^2 only after the base Hamiltonian loss is stable. The paper supplies the contact-set characterization, while the velocity stencil, finite horizon, grid interpolation, and tolerance are empirical approximations. Test first on a 1D or 2D periodic mechanical Hamiltonian with a reference Aubry set from a high-resolution solver. Compare uniform sampling with contact-guided sampling at equal H evaluations. Measure Aubry-set Hausdorff distance, dense-grid Hamiltonian violation, and value error. Success is faster error reduction near separatrices and improved contact-set localization at equal sample count.

## Disclaimer

AI-generated research hypothesis, automatically tested. Not peer-reviewed.