{
 "artifacts": null,
 "category": "dynamics",
 "description": "Replace the assumption that strong convexity keeps optimization inside a valid parameter chart with an explicit viability condition on the chart boundary. For Lie-group neural-network parameters or bounded latent coordinates, modify each update so its velocity has nonpositive outward radial component, using either a radial barrier or projection onto the tangent cone.",
 "formulas_latex": [
  "$$f(x)=\\frac{1}{2}(x-x^*)^\\top H(x-x^*),\\qquad H=\\begin{bmatrix}1\u00260.7\u00260\\\\0.7\u00261\u00260\\\\0\u00260\u00261\\end{bmatrix},\\qquad x^*=\\begin{pmatrix}0.23\\\\0.16\\\\0\\end{pmatrix}.$$",
  "$$\\mathrm{Hess}_{\\mathrm{SO}(3)}f\\succeq0.172I\\ \\text{on}\\ \\|x\\|\\leq0.3,\\qquad \\left.\\frac{d}{dt}\\|x\\|\\right|_{\\|x\\|=0.3}=\\frac{21}{500}\u003e0\\ \\text{at the certified boundary point}.$$",
  "$$\\frac{d}{dt}\\|x\\|^2=2x^\\top v(x)\\leq0\\quad(\\|x\\|=\\rho),\\qquad v_{\\mathrm{barrier}}(x)=-\\nabla f(x)-\\lambda_B\\nabla B(x),\\quad B(x)=-\\log\\left(1-\\frac{\\|x\\|^2}{\\rho^2}\\right).$$",
  "$$v_{\\mathrm{proj}}(x)=v(x)-\\max\\left(0,\\frac{x^\\top v(x)}{\\|x\\|^2}\\right)x.$$"
 ],
 "id": 2800,
 "implementation": "1. Integration point: apply this to any neural module whose parameters are represented in a Lie-algebra chart, such as rotation-valued layers, equivariant networks, neural ODE states constrained to a ball, or a low-dimensional latent code with a required trust region. Compute the ordinary minibatch gradient velocity \\(v=-g\\), then modify it before the optimizer update. Start with the cheap projection rule; add the barrier as a differentiable alternative. For a discrete step, use \\(x_{k+1}=x_k+\\eta v_{\\mathrm{proj}}(x_k)\\), or retract a Lie-group parameter with \\(R_{k+1}=R_k\\exp(\\eta\\widehat v_{\\mathrm{proj}})\\). 2. Pseudocode: compute loss and gradient \\(g\\); set \\(v=-g\\); if \\(\\|x\\|\\geq\\rho-\\epsilon\\), compute \\(a=x^\\top v/\\|x\\|^2\\) and replace \\(v\\leftarrow v-\\max(0,a)x\\); update using the exponential-map retraction. For the barrier version, use \\(v\\leftarrow-g-\\lambda_B 2x/[\\rho^2(1-\\|x\\|^2/\\rho^2)]\\). 3. Computed versus estimated: \\(x\\), its norm, and the boundary projection are computed exactly per batch; the paper's curvature certificate is not needed by the implementation, while the actual outward-rate distribution is estimated by logging \\(x^\\top v/\\|x\\|\\) on near-boundary iterates. 4. First experiment: train a small rotation-equivariant MLP or SO(3)-parameterized linear layer on synthetic rotated-MNIST, comparing Adam, unconstrained SGD, projected flow, and barrier flow with \\(\\rho=0.3\\). Initialize near the paper's quadratic example and use the same \\(H\\) as a controlled test. The falsifiable signature is that the unconstrained method exhibits positive boundary-normal velocity near the boundary, with a controlled reproduction near \\(0.042\\), and exits the chart; projection should make every logged boundary-normal velocity nonpositive up to numerical tolerance and produce zero chart exits. In discrete time, measure the largest safe learning rate and test whether it scales approximately as the distance-to-boundary divided by the maximum outward velocity.",
 "math_summary": "Let \\(x=\\log R\\in\\mathbb{R}^{3}\\), \\(\\mathcal C_\\rho=\\{x:\\|x\\|\\leq\\rho\\}\\), and \\(f(x)=\\tfrac12(x-x^*)^\\top H(x-x^*)\\), with \\(H=\\begin{bmatrix}1\u00260.7\u00260\\\\0.7\u00261\u00260\\\\0\u00260\u00261\\end{bmatrix}\\), \\(x^*=(0.23,0.16,0)^\\top\\), and \\(\\rho=0.3\\). The paper certifies \\(\\mathrm{Hess}_{\\mathrm{SO}(3)}f\\succeq0.172I\\) throughout the ball, yet the gradient flow has an exact outward radial velocity \\(21/500=0.042\\) at a boundary point. For a Euclidean chart, forward invariance requires \\(\\frac{d}{dt}\\|x\\|^2=2x^\\top v(x)\\leq0\\) whenever \\(\\|x\\|=\\rho\\), where \\(v\\) is the parameter velocity. A differentiable barrier \\(B(x)=-\\log(1-\\|x\\|^2/\\rho^2)\\) gives \\(\\nabla B(x)=2x/[\\rho^2(1-\\|x\\|^2/\\rho^2)]\\), whose coefficient diverges at the boundary and suppresses outward motion.",
 "math_tags": [
  "dynamical-systems",
  "differential-geometry",
  "control-theory",
  "optimization"
 ],
 "ml_areas": [
  "optimizer",
  "training-dynamics",
  "regularization"
 ],
 "paper": {
  "arxiv_id": "2608.28976",
  "arxiv_url": "https://arxiv.org/abs/2608.28976",
  "summary_what_math_gives_to_ml": "The paper provides a certified counterexample showing that geodesic strong convexity and an interior minimizer do not imply forward invariance of a geodesic trust region under gradient flow on \\(\\mathrm{SO}(3)\\). Its transferable mechanism is a separation between curvature relative to the minimizer and the boundary-normal component required by the viability condition: a strongly convex objective can still point outward at the boundary. Neural-network training on Lie-group parameters, latent rotations, or any bounded chart should therefore enforce a boundary-normal condition or use a barrier/projected flow rather than relying on strong convexity alone. The sharp engineering prediction is that unconstrained updates leave the chart at a measurable step size, while a radial barrier or projection eliminates outward boundary velocity.",
  "title": "Geodesic strong convexity does not imply forward invariance under gradient flow on SO(3): a certified counterexample",
  "year": "2026"
 },
 "ratings": {
  "difficulty": 4,
  "novelty": 6,
  "usefulness": 6
 },
 "solves": [
  "stability",
  "training"
 ],
 "title": "Boundary-normal trust-region flow",
 "url": "https://synthcore.org/idea/2800/boundary-normal-trust-region-flow",
 "verification": {
  "peer_reviewed": false,
  "status": "unverified",
  "status_label": "Unverified",
  "verdict_source": "deterministic test code (paired-seed permutation statistics)"
 }
}