Solves: Stability

Machine-learning ideas tagged Stability in the Solves taxonomy of the Math2NN corpus.

2414 ideas found

Unverified 2026

Cosine-Covered Flat-Region Escape

Augment gradient descent with a directional-search step when the gradient norm is small or the loss has stalled. In each parameter block, evaluate a small positively spanning set of normalized perturbations, use their directional loss slopes to identify descent directions, and combine them through nonnegative coefficients so that the update remains inside their positive span. The cosine measure supplies a quantitative trigger: low directional coverage means the current perturbation pool is not…

Useful5/10
Difficulty5/10
Novelty6/10
Paper: The cosine measure of a function at a point arXiv:2608.07716
Unverified 2026

Spread-complexity spectral regularizer

Regularize the eigenvalue spectrum of a neural representation or attention Gram matrix using the paper's universal-kernel spread-complexity curve. The loss penalizes spectral profiles that exhibit excessive level clustering or near-degeneracy, while allowing the desired amount of eigenvalue repulsion to be selected by a GOE-like, Poisson-like, or empirically calibrated target.

Useful5/10
Difficulty5/10
Novelty7/10
Paper: Analytic Spread Complexity from Level Statistics: From Chaos to Integrability arXiv:2608.07412
Unverified 2026

Worst-Subset Conditioning Regularizer

Train an overcomplete linear or MLP layer so that square subsets of its output rows remain numerically invertible after neuron pruning or routing failures. Penalize sampled subsets with unusually small least singular values, using the paper's entropy exponent to quantify the severity expected from random redundancy.

Useful5/10
Difficulty5/10
Novelty7/10
Paper: Extreme least singular values of random row submatrices with bounded-density subgaussian entries arXiv:2608.07410
Unverified 2026

Logarithmic-Laplacian Feature Regularizer

Add a nonlocal logarithmic-Laplacian penalty to intermediate spatial feature maps or ordered token embeddings. Unlike a standard graph or image Laplacian, the kernel uses scale-free weights proportional to |z|^{-n} and includes a local compensation term, allowing multiscale feature smoothing without simply forcing nearby features to become identical.

Useful5/10
Difficulty5/10
Novelty7/10
Paper: Hölder regularity and Harnack inequality for the logarithmic Laplacian arXiv:2608.07315
Unverified 2026

Mobility-matched entropy sampler

Build a neural sampler whose deterministic probability-flow dynamics implement the nonlinear Fokker–Planck equation rather than the usual linear Langevin flow. For a selected monotone diffusion law \(P\), use the associated entropy derivative \(\phi'(r)=P'(r)/r\) to define the chemical potential and train a neural velocity field to approximate its descent direction.

Useful5/10
Difficulty6/10
Novelty6/10
Paper: Nonlinear Diffusion Equations: Full characterization of Entropies arXiv:2608.07129
Unverified 2026

Global-Shift Randomized Batching

Sample one batching aggressiveness parameter once per serving epoch and use it for all busy periods, instead of independently randomizing every batch. This globally coupled randomness preserves nested batching behavior while preventing a fixed traffic pattern from repeatedly aligning with a deterministic timeout.

Useful5/10
Difficulty3/10
Novelty7/10
Paper: Online Multi-Level Aggregation with Per-Batch Maximum Delay arXiv:2608.06796
Unverified 2026

Vacancy-preserving collision-free router

Build a differentiable assignment layer whose rows represent tokens and whose columns represent experts, memory slots, or attention slots. Each row has unit probability mass, but no column receives positive mass from two rows; maintaining at least one vacant column makes assignments continuously deformable through elementary vacancy moves instead of abrupt softmax switches.

Useful5/10
Difficulty6/10
Novelty5/10
Paper: Tilings, packings, and the existence of Schwartz-class Gabor windows arXiv:2608.06679
Unverified 2026

Active-Contact Cone Projection

Replace soft pairwise repulsion between learned prototypes or codebook vectors with an active-set feasibility layer based on the paper's first-order admissible cone. Pairs exactly at the minimum distance contribute linear half-space constraints to the update, while separated pairs do not unnecessarily restrict motion. This should reduce prototype collapse and make constrained embedding or quantization training less sensitive to penalty weights.

Useful5/10
Difficulty5/10
Novelty4/10
Paper: Intrinsic Geometry of Hard Disk Clusters arXiv:2608.06513
Unverified 2026

Vineyard Activation Monitor

Construct a filtered cell complex from neural activations or a learned token/feature graph and track its persistence barcode incrementally as model activations change. Replace full persistent-homology recomputation at every checkpoint by maintaining homology bases and applying local transpositions when filtration blocks split or merge; use barcode drift as a training monitor or a weak regularization signal.

Useful5/10
Difficulty6/10
Novelty6/10
Paper: Computing Conley-Morse Persistence Barcode Efficiently by Updating Matrix Decompositions arXiv:2608.06507
Unverified 2026

Sharp independent-load tail regularizer

Apply the paper's extremal tail bound to independently sampled nonnegative neural-network contributions, such as stochastic-depth branch activations, independently gated expert loads, or separately allocated memory chunks. Penalize the analytic worst-case probability that their sum exceeds a budget, using the fact that the worst admissible distribution is a sparse Bernoulli spike at the threshold.

Useful5/10
Difficulty4/10
Novelty8/10
Paper: Sharp Tail Bounds Beyond Twice the Mean arXiv:2608.06317
Unverified 2026

Product-observability regularizer

Represent cross-modal or two-stream interactions as a bipartite tensor and explicitly maximize their response to product observables rather than allowing all information to be hidden in inseparable global interactions. Penalize interactions whose global trace norm is large but whose best product-observable response is small, using the paper's sharp bound as a dimension-aware calibration.

Useful5/10
Difficulty6/10
Novelty6/10
Paper: Global vs. Product Observables in Bipartite Quantum Systems: The Sharp Bound arXiv:2608.06235
Unverified 2026

Profile Consistency Regularizer

Regularize an encoder so that geometrically equivalent augmentations preserve the colored interaction profile across scales. Unlike a scalar overlap loss, the objective penalizes changes in connected overlap and alternating higher-dimensional topology simultaneously over a radius grid.

Useful5/10
Difficulty4/10
Novelty6/10
Paper: The Intersection Euler Characteristic Profile: Euler Calculus and Stability for Topological Interaction of Ball Unions arXiv:2608.06180
Unverified 2026

Bartlett-LKJ Correlated Head Noise

Replace independent dropout or Gaussian perturbations across attention heads, ensemble members, or diffusion score replicas with a positive-semidefinite correlation matrix sampled from an LKJ distribution. The concentration parameter eta controls whether perturbations are nearly independent or strongly correlated in a controlled way, while the Bartlett construction guarantees a valid covariance without matrix rejection or projection.

Useful5/10
Difficulty4/10
Novelty7/10
Paper: Bartlett Couplings of the Onion and Vine LKJ Samplers arXiv:2608.06116
Unverified 2026

Auxiliary-energy neural optimizer

Replace the direct nonlinear loss step by a scalar-auxiliary-variable discretization of a gradient flow. The optimizer maintains an auxiliary value representing the square root of the nonlinear energy, so the coupled update has a discrete modified-energy decrease even when the step size is not restricted by the local curvature of the loss.

Useful5/10
Difficulty6/10
Novelty7/10
Paper: A Thermodynamically Consistent Cahn-Hilliard-Navier-Stokes Model for Tumor Growth arXiv:2608.06099
Unverified 2026

Square-Root Boundary-Temperature Attention

Add a measurement-conditioned attention layer with two explicitly separated fields: a geometry-only inverse-temperature profile that controls interaction strength and an outcome-dependent chemical-potential bias. For a region bounded by coordinates a and b, force the interaction gate to vanish as the square root of the distance from either boundary, while allowing a separate potential channel to encode measured values.

Useful5/10
Difficulty4/10
Novelty7/10
Paper: Measurement-induced entanglement Hamiltonian arXiv:2608.06006
Unverified 2026

Lyapunov-Continuation Initialization for Noisy State-Space Layers

Initialize and train a linear recurrent or state-space transition using the stochastic Lyapunov operator rather than only constraining the drift matrix to be Hurwitz. Start from a controller that stabilizes the drift-only dynamics, then continuously increase the multiplicative-noise coefficient and update the controller while enforcing a positive-definite Lyapunov certificate. The resulting module should avoid exploding hidden states when process noise depends on the hidden state or input.

Useful5/10
Difficulty5/10
Novelty6/10
Paper: Stabilizer Design for Policy Iteration in Stochastic Linear Quadratic Control: A Spectrum-Assignment Approach arXiv:2608.05953
Unverified 2026

Dynamic-scaling cyclic optimizer

Drive the optimizer periodically around a baseline learning rate, but scale the modulation amplitude and period through a single dimensionless control variable rather than tuning them independently. The neural analogue predicts that normalized loss, gradient norm, and parameter-displacement trajectories should approximately collapse across schedules with equal \(aP^{\kappa}\), while sufficiently large values should reveal a measurable transition from weak tracking to strongly oscillatory or…

Useful5/10
Difficulty4/10
Novelty8/10
Paper: Dynamic scaling behavior in the presence of a periodic magnetic driving across Ising continuous transitions arXiv:2608.05936
Unverified 2026

Mass-Covering Dimension Regularizer

Regularize hidden representations using the number of metric balls required to cover at least a fixed fraction of minibatch probability mass. The outlier tolerance ignores a controlled fraction of atypical samples, while the resolution parameter makes the penalty explicitly scale-dependent. Combine the penalty with a variance floor or reconstruction term so that reducing geometric dimension does not produce a constant representation.

Useful5/10
Difficulty5/10
Novelty6/10
Paper: Complexity and Stability of Neural Activity Across Aging and Neurodegenerative Disease arXiv:2608.05882
Unverified 2026

Splitting-Conjugate Latent Dynamics

Equip a latent transition model with a near-identity polynomial coordinate transform that conjugates the nonlinear transition to a linear latent operator, at least locally around a reference state. Train the transform jointly with the dynamics using both the usual prediction loss and the paper's splitting/intertwining residual, so that multi-step prediction is performed partly in approximately linearised coordinates.

Useful5/10
Difficulty6/10
Novelty6/10
Paper: Linearisation, splitting property and homotopy algebras arXiv:2608.05875
Unverified 2026

Spin-Wave Nonlinearity Damping

Use the paper's exponential dressing of an activity coupling as an adaptive gate on a neural network's nonlinear residual branch. The branch is strongly suppressed when the local activation fluctuation variance is high, producing an automatically linearized and more stable update, while low-variance representations preserve the learned nonlinear interaction.

Useful5/10
Difficulty3/10
Novelty6/10
Paper: Large Spin-Wave Fluctuations Suppress Activity in Malthusian Flocks arXiv:2608.05805
Unverified 2026

Gaussian Minkowski Concavity Regularizer

Represent each class or concept by a convex latent body containing the origin, and penalize violations of the paper's sharp Gaussian Brunn–Minkowski inequality when two bodies are interpolated by Minkowski addition. This regularizes latent supports toward geometries whose Gaussian probability mass remains predictable under interpolation, potentially improving interpolation robustness and out-of-distribution behavior.

Useful5/10
Difficulty7/10
Novelty8/10
Paper: The Brunn--Minkowski inequality for the Gaussian measure arXiv:2608.05390
Unverified 2026

Fuzzy permutation attention

Replace part of an attention matrix with a mixture of fuzzy permutation matrices induced by short permutations. Each basis element represents an order-preserving k-token matching smeared over all embeddings into the sequence, while a balancing constraint makes the aggregate attention receive uniform global coverage. Retain a standard low-rank or local-attention residual so the structured branch does not prevent arbitrary content-dependent interactions.

Useful5/10
Difficulty5/10
Novelty7/10
Paper: Fuzzy latin squares and balanced permutation pattern statistics arXiv:2608.05335
Unverified 2026

Chain-Compatible Differential Pooling

Treat learned features on a mesh as differential forms and pool them against oriented chains using wedge or cap products instead of ordinary coordinate averaging. Couple forward and boundary features with the signed chain differential so that pooling commutes with differentiation, preserving local conservation and orientation information.

Useful5/10
Difficulty5/10
Novelty8/10
Paper: Differential Homology arXiv:2608.05048
Unverified 2026

Frame-safe totally-positive front-end

Replace the first learned one-dimensional convolution or STFT-like feature extractor with a differentiable bank of time-frequency shifts of a totally positive window. Parameterize the temporal spacing \(\alpha\) and frequency spacing \(\beta\) so that \(\alpha\beta<1\) is always satisfied, giving a mathematically certified oversampled representation instead of an arbitrarily subsampled filterbank.

Useful5/10
Difficulty5/10
Novelty5/10
Paper: Gabor Frames of Totally Positive Functions: A Complete Characterization arXiv:2608.04992