✗ Mechanism failed
2026
Replace the random or gradient-aligned perturbation in sharpness-aware minimization with a unit perturbation direction selected by a polynomial of the local Hessian. With \(\mathscr{P}(s)=(s-\rho)^2\), the direction converges toward Hessian eigenspaces whose eigenvalues are closest to the target curvature \(\rho\), allowing regularization of a chosen curvature band instead of indiscriminately penalizing only the sharpest direction.
Useful6/10
Difficulty6/10
Novelty7/10
✗ Mechanism failed
2026
Replace scalar entropy penalties on attention maps with a matrix-valued heat-flow regularizer over a circular or periodic token coordinate. Each position stores a positive semidefinite matrix describing coupled heads, experts, or channels; heat smoothing is constrained by the sharp modified log-Sobolev and Bogoliubov–Kubo–Mori contraction rather than an arbitrary smoothing coefficient. This should suppress high-frequency routing noise while preserving positive matrix structure and reducing…
Useful6/10
Difficulty6/10
Novelty7/10
△ Mechanism confirmed, baseline not beaten
2026
Replace a contractive recurrent transition by its explicit Schäffer isometric lift, optionally augmenting it with a second operator satisfying the nonlinear covariance relation $V_1V_2=V_2f(V_1)$. The lifted state preserves or nearly preserves hidden-state energy, while the covariance penalty or parameterization imposes an algebraic structure on multiple recurrent channels.
Useful6/10
Difficulty6/10
Novelty6/10
△ Mechanism confirmed, baseline not beaten
2026
Replace ordinary Jacobian penalties in coordinate MLPs or deformation networks with a learned local rotation frame and a polyconvex energy of the relative stretch. Penalize \(U\), its cofactor, and its determinant through a convex function, while separately smoothing the rotation field through \(R^T\operatorname{Curl}R\). The intended benefit is resistance to fold formation and better conditioning than directly penalizing \(\|J-I\|^2\), especially for large deformations.
Useful6/10
Difficulty6/10
Novelty6/10
✓✓ Beats tuned baseline
2026
Replace one deterministic residual update with a short cyclic composition of learned vector fields evaluated for randomized, short run times. Because finite compositions of noncommuting flows generate directional-derivative and Lie-bracket terms, changing the cycle order gives the network an explicit, low-cost way to learn drift directions that are unavailable from the individual vector fields alone.
Useful6/10
Difficulty5/10
Novelty7/10
✗ Failed on benchmark
2026
Add a loss term requiring a neural optimizer or recurrent module to decrease a nonnegative Lyapunov-like energy over M update steps, rather than forcing monotonic one-step decrease. The term includes an empirically estimated mismatch allowance, so stochastic or delayed updates are tolerated while persistent instability remains penalized.
Useful6/10
Difficulty4/10
Novelty7/10
✗ Mechanism failed
2026
Treat a recurrent or state-space layer as a finite-state Markov cocycle and constrain optimizer steps using the paper's inverse-logarithmic sensitivity of Lyapunov exponents near a zero exponent gap. Instead of enforcing a crude spectral-norm bound, allow updates that are harmless for long-run growth while shrinking steps that could substantially change the recurrent stability profile.
Useful6/10
Difficulty5/10
Novelty7/10
✗ Failed on benchmark
2026
Regularize a circular recurrent kernel by directly controlling the growth rate and phase velocity of its Fourier modes. This converts replay-speed selection into a low-dimensional spectral control problem and can suppress unstable or excessively slow modes without adding recurrent parameters.
Useful6/10
Difficulty6/10
Novelty8/10
△ Mechanism confirmed, baseline not beaten
2025
Regularize a neural dynamical map so that its log-volume expansion is cohomologous to a constant rather than forcing the Jacobian determinant to be constant at every state. Learn a scalar potential that explains transient expansion and penalize only the non-telescoping component, which should reduce long-horizon gradient explosion or collapse while retaining useful average expansion.
Useful6/10
Difficulty5/10
Novelty7/10
Unverified
2026
Add asynchronous nonlinear consensus steps between local-SGD updates in a federated or decentralized system. Instead of averaging a participating client with all peers, replace one client parameter vector by the coordinatewise p-mean of neighboring replicas, allowing sparse communication and reduced sensitivity to atypical client models.
Useful5/10
Difficulty5/10
Novelty7/10
Unverified
2026
Use the two-species phase diagram as a training controller: gradually increase one-way mutation \(\lambda\) and mutant suppression \(\mu\) while monitoring whether feature activity approaches the absorbing boundary. This creates a measurable curriculum from independent learning to controlled competition instead of turning on strong destructive interactions at initialization.
Useful5/10
Difficulty4/10
Novelty8/10
Unverified
2026
Use state-dependent Langevin noise to perturb prototype or router-codebook optimization when balanced quantization reaches a poor stationary configuration. Retain the best-so-far state for deployment, allowing exploration to escape non-minimizing Lloyd fixed points without permanently corrupting the learned representation.
Useful5/10
Difficulty4/10
Novelty5/10
Unverified
2026
Use diagonal similarity scaling to reduce the induced infinity norm of a sparse recurrent transition or companion operator before or during training. Similarity preserves eigenvalues while changing coordinate-wise amplification, so the method can suppress non-normal transient growth without changing the represented linear dynamics.
Useful5/10
Difficulty5/10
Novelty6/10
Unverified
2026
Replace a learned linear operator by Monte Carlo averages of unitary-conjugated copies, where the conjugators are discrete time-frequency shifts sampled from a narrow phase-space distribution. The conjugation preserves the singular spectrum of each sampled operator, while the expectation penalizes sensitivity to small translations and modulations without directly shrinking the operator norm.
Useful5/10
Difficulty5/10
Novelty7/10
Unverified
2026
Use the reciprocal arrangement as a probe of whether a learned representation has the intended angular response, and penalize deviations from the paper's universal beta distribution. This converts the theorem into a distribution-level regularizer rather than assuming that the reciprocal layer itself improves task loss.
Useful5/10
Difficulty4/10
Novelty7/10
Unverified
2026
Replace a static top-k MoE capacity rule with a router whose expert allocation evolves through a finite-domain coverage process. Experts with larger current occupancy can either receive more future capacity, intentionally amplifying specialization, or receive less capacity by reversing the size dependence, allowing a controlled test of the paper's asymmetry-amplification mechanism.
Useful5/10
Difficulty6/10
Novelty6/10
Unverified
2026
Represent a neural model's particle ensemble, latent samples, or routing prototypes as an empirical probability measure and penalize its Wasserstein total variation across training or inference steps. Discrete resampling and particle replacement remain allowed, but their mass-distance cost is made explicit so the model cannot obtain a cheap distributional change through untracked teleportation. A weak continuity-equation residual can be added as an auxiliary loss or used as a diagnostic.
Useful5/10
Difficulty6/10
Novelty6/10
Unverified
2026
Train two parameter replicas with common low-rank stochastic forcing and an adaptive finite-dimensional Cameron–Martin correction that contracts their discrepancy in a weak parameter metric. Transporting the forcing directions through the loss Hessian is intended to make a rank-k perturbation influence more than k raw parameter directions, while damped momentum suppresses high-energy divergence.
Useful5/10
Difficulty7/10
Novelty7/10
Unverified
2026
Constrain a channel-mixing layer to be a product of nonnegative bidiagonal matrices, rather than an unconstrained dense matrix. The resulting totally nonnegative operator is predicted not to increase sign oscillations in ordered channel features, potentially reducing high-frequency feature noise and making deep stacks more stable.
Useful5/10
Difficulty4/10
Novelty8/10
Unverified
2026
Add a differentiable penalty that encourages a neural implicit field to have a controlled local homogeneity degree across concentric spatial scales. The penalty compares the flux-normalized frequency at adjacent radii, optionally targeting a desired degree k, so the network is discouraged from producing scale-inconsistent or oscillatory local geometry.
Useful5/10
Difficulty5/10
Novelty7/10
Unverified
2026
Replace two unconstrained scalar quadratic feature heads with a pair whose quadratic forms admit a positive-definite linear combination. This prevents the two heads from simultaneously vanishing on any nonzero hidden vector, which can reduce representation collapse and improve the conditioning of downstream gates or auxiliary objectives. The constraint can be implemented softly with a spectral-margin penalty, or exactly by parameterizing one learned pencil as positive definite.
Useful5/10
Difficulty4/10
Novelty8/10
Unverified
2026
Use the Gaussian approximation of a high-dimensional maximum to set a simultaneous coordinate-clipping threshold for minibatch gradients or activations. The threshold is sampled from a correlated Gaussian with the observed batch covariance, rather than treating coordinates as independent or estimating an unstable extreme quantile directly.
Useful5/10
Difficulty6/10
Novelty7/10
Unverified
2026
Replace continuous stochastic-gradient updates by a flashing schedule with alternating ON phases, where gradients act normally, and OFF phases, where gradients are suppressed or weakened and controlled noise allows escape from local traps. Estimate directional asymmetry of the local loss basin from forward and backward probe distances, then set the flashing frequency using the ratchet resonance law so that noise-assisted transitions preferentially produce net progress toward lower loss.
Useful5/10
Difficulty6/10
Novelty8/10
Unverified
2026
Treat the number K of minibatches between expensive control updates as a review period: the controlled neural dynamics use parameters or decisions computed at time nK and hold them fixed until (n+1)K. Scan K, estimate first and second finite differences of validation loss or episodic return, and use the resulting nonmonotone-to-convex or concave phase diagram to select an update frequency rather than assuming that more frequent updates are always better.
Useful5/10
Difficulty4/10
Novelty6/10