{
 "artifacts": null,
 "category": "optimization",
 "description": "Replace decentralized parameter averaging with consensus on cumulative local gradient states, followed by a barrier-FTRL update that stays strictly inside a convex feasible set. This is particularly suitable for federated learning with heterogeneous clients and for simplex-constrained mixture, router, or adapter parameters, where Euclidean projection can be unstable or expensive.",
 "formulas_latex": [
  "$$\\mathfrak{R}_{\\alpha}^{a}(T):=\\mathbb{E}\\left[\\alpha\\max_{x\\in\\mathcal{K}}\\sum_{t=1}^{T}\\mathsf{F}_{t}(x)-\\sum_{t=1}^{T}\\mathsf{F}_{t}(x_{a,t})\\right],\\qquad \\mathsf{F}_t(x)=\\frac{1}{N}\\sum_{j=1}^{N}f_{t,j}(x).$$",
  "$$y_{i,t+1}=\\sum_{j=1}^{N}W_{ij}y_{j,t}+g_{i,t},\\qquad g_{i,t}=\\nabla\\ell_{i,t}(x_{i,t}),\\qquad x_{i,t}=\\arg\\min_{x\\in\\operatorname{int}(\\mathcal{K})}\\left\\{\\langle y_{i,t},x\\rangle+\\frac{1}{\\eta}\\psi(x)\\right\\}.$$",
  "$$y_{i,t}+\\frac{1}{\\eta}\\nabla\\psi(x_{i,t})=0,\\qquad \\Delta x=-\\left(\\frac{1}{\\eta}\\nabla^2\\psi(x)\\right)^{-1}\\left(y_{i,t}+\\frac{1}{\\eta}\\nabla\\psi(x)\\right),\\qquad x\\leftarrow x+\\tau\\Delta x,$$",
  "$$\\psi(x)=-\\sum_{k=1}^{d}\\log x_k,\\qquad \\nabla\\psi(x)_k=-x_k^{-1},\\qquad \\nabla^2\\psi(x)=\\operatorname{diag}(x_1^{-2},\\ldots,x_d^{-2})$$"
 ],
 "id": 2882,
 "implementation": "(1) Exact integration point: implement this as a decentralized optimizer for a tensor constrained to a convex set. Start with a learnable MoE router probability vector or adapter-mixture vector $x_i\\in\\Delta^{d-1}$ on each client; ordinary network weights can remain AdamW-trained, while the selected simplex tensor is updated by Barrier-FTRL. Each client stores only a dual vector $y_i\\in\\mathbb{R}^d$ and communicates that vector to graph neighbors once per training round.\n\n(2) Pseudocode:\n```\ninitialize y_i = 0 and x_i = uniform_simplex(d)\nfor t in 1..T:\n    x_i = NewtonSolve(y_i, psi, eta)\n    loss_i = minibatch_loss(client_i, x_i, other_parameters)\n    g_i = grad_x(loss_i, x_i)\n    receive y_j from graph neighbors\n    y_i = sum_j W[i,j] * y_j + g_i\n    x_i = NewtonSolve(y_i, psi, eta)\n```\nInside `NewtonSolve`, evaluate $r=y_i+\\nabla\\psi(x)/\\eta$, $H=\\nabla^2\\psi(x)/\\eta$, compute $\\Delta x=-H^{-1}r$, choose backtracking $\\tau$ so $x+\\tau\\Delta x$ remains in the barrier domain and decreases $\\langle y_i,x\\rangle+\\psi(x)/\\eta$, and stop when $\\|r\\|_\\infty\u003c10^{-6}$. For the simplex, use reduced coordinates with $x_d=1-\\sum_{k\u003cd}x_k$, or solve directly with a Lagrange multiplier enforcing $\\sum_kx_k=1$.\n\n(3) Computed from the paper's mathematics: the barrier objective, gradient, Hessian, Newton decrement, and consensus update with a symmetric doubly stochastic matrix $W$. Estimate empirically: choose $\\eta$ by a logarithmic sweep, measure the second-largest eigenvalue magnitude $\\rho=|\\lambda_2(W)|$ to characterize mixing, and compare one-neighbor communication against FedAvg and decentralized SGD. If client gradients are noisy, use $g_i$ as the minibatch gradient and optionally clip only the dual increment before mixing; do not clip the primal vector, because feasibility is supplied by the barrier.\n\n(4) First cheap experiment: train a small CIFAR-10 classifier with four simulated non-IID clients, using a four-component simplex adapter or router vector. Compare Barrier-FTRL, decentralized SGD, and FedAvg at equal client minibatch FLOPs and equal communicated bytes. Record training loss, validation accuracy, simplex-boundary violations, gradient disagreement $N^{-1}\\sum_i\\|y_i-\\bar y\\|^2$, and communication rounds. The first success signal is smoother loss under heterogeneous clients, no zero-probability router collapse, and higher validation accuracy at the same communication budget; a useful target is 10-20% fewer rounds to reach a fixed accuracy or a statistically significant reduction in client-to-client loss variance.",
 "math_summary": "The paper measures network-aggregate alpha-regret by $$\\mathfrak{R}_{\\alpha}^{a}(T):=\\mathbb{E}\\left[\\alpha\\max_{x\\in\\mathcal{K}}\\sum_{t=1}^{T}\\mathsf{F}_{t}(x)-\\sum_{t=1}^{T}\\mathsf{F}_{t}(x_{a,t})\\right],$$ where $N$ is the number of agents, $a$ is an agent, $T$ is the number of rounds, $\\mathcal{K}$ is a compact convex feasible set, $f_{t,j}$ is agent $j$'s payoff at round $t$, and $\\mathsf{F}_t(x)=N^{-1}\\sum_{j=1}^N f_{t,j}(x)$ is the network-average payoff. The implementable minimization analogue maintains a dual cumulative-gradient state $y_{i,t}$ at worker $i$ and chooses the primal iterate by barrier FTRL: $x_{i,t}=\\arg\\min_{x\\in\\operatorname{int}(\\mathcal{K})}\\{\\langle y_{i,t},x\\rangle+\\eta^{-1}\\psi(x)\\}$, where $\\eta\u003e0$ is the step size and $\\psi$ is a strictly convex barrier diverging at the boundary. Communication uses a symmetric doubly stochastic mixing matrix $W$: $y_{i,t+1}=\\sum_{j=1}^{N}W_{ij}y_{j,t}+g_{i,t}$, where $g_{i,t}=\\nabla\\ell_{i,t}(x_{i,t})$ is the local loss gradient. The barrier update is characterized by $y_{i,t}+\\eta^{-1}\\nabla\\psi(x_{i,t})=0$ and can be solved by damped Newton iterations with Hessian $H=\\eta^{-1}\\nabla^2\\psi(x)$. For a simplex $\\mathcal{K}=\\{x\\in\\mathbb{R}^d:x_k\u003e0,\\sum_kx_k=1\\}$, use the log barrier $\\psi(x)=-\\sum_{k=1}^d\\log x_k$; its Hessian is diagonal, $\\nabla^2\\psi(x)=\\operatorname{diag}(x_k^{-2})$, so Newton solves are cheap. The useful property is that iterates remain interior and the state tracks the network-average cumulative gradient despite only one neighbor exchange per round.",
 "math_tags": [
  "optimization",
  "convex-analysis",
  "numerical-analysis",
  "linear-algebra"
 ],
 "ml_areas": [
  "optimizer",
  "federated",
  "moe-routing"
 ],
 "paper": {
  "arxiv_id": "2608.30271",
  "arxiv_url": "https://arxiv.org/abs/2608.30271",
  "summary_what_math_gives_to_ml": "The paper's transferable contribution is a decentralized online optimizer that keeps each worker's communication state in dual space, rather than repeatedly exchanging model parameters or full historical gradients. Barrier-FTRL combines cumulative surrogate gradients with a self-concordant barrier, producing feasible iterates even when the action set is accessed only through separation queries; approximate Newton solves make the update computationally practical. For neural networks, the most direct test is a constrained federated optimizer for simplex, box, or norm-bounded parameters, where local clients mix dual gradient states and solve a barrier-regularized proximal subproblem. The expected benefit is improved stability under heterogeneous client objectives and exact feasibility without projection clipping.",
  "title": "Dec-BFTRL: Squre-Root Regret for Decentralized Online Upper-Linearizable Optimization under Separation Access with Application to Continuous Submodular Maximization",
  "year": "2026"
 },
 "ratings": {
  "difficulty": 5,
  "novelty": 6,
  "usefulness": 6
 },
 "solves": [
  "stability",
  "scalability",
  "accuracy"
 ],
 "title": "Decentralized Barrier-FTRL Optimizer",
 "url": "https://synthcore.org/idea/2882/decentralized-barrier-ftrl-optimizer",
 "verification": {
  "peer_reviewed": false,
  "status": "unverified",
  "status_label": "Unverified",
  "verdict_source": "deterministic test code (paired-seed permutation statistics)"
 }
}