✗ Failed on benchmark
2026
Replace ordinary gradient descent in a chosen approximately linear parameter block with gradient descent plus a controlled negative quadratic penalty, and stop before the unstable directions explode. The finite-time spectral filter can amplify well-supported directions while retaining shrinkage or limited exposure on weak directions, which is unavailable to a stable negative-ridge endpoint.
Useful7/10
Difficulty5/10
Novelty7/10
✓✓ Beats tuned baseline
2026
Train a mixture-of-experts router by solving its regularized nonnegative simplex least-squares subproblem with a matrix-free active-set conjugate-gradient method instead of projected gradient or Adam. The router coefficients remain exactly nonnegative and sum to one, while CG rapidly solves each free-set quadratic and the active-set pivots identify sparse expert assignments.
Useful7/10
Difficulty5/10
Novelty7/10
△ Mechanism confirmed, baseline not beaten
2026
Track a symmetry-asymmetry functional of network outputs or hidden states under a chosen transformation or channel and estimate which relaxation modes control its late-time decay. Use a short warm-up trajectory to suppress the slow asymmetry mode while allowing larger initial asymmetry in faster modes, producing a training trajectory that can overtake a nominally better-initialized trajectory. This transfers the paper's quantum Mpemba effect as a mode-overlap principle rather than requiring an…
Useful7/10
Difficulty6/10
Novelty8/10
✗ Failed on benchmark
2026
Replace the usual parameter-plus-momentum Langevin state with a three-level chain consisting of parameters, velocity, and acceleration, while injecting Gaussian noise only into the highest auxiliary state. At a saddle, the escaping direction has a positive rate given by a cubic characteristic equation; use this rate to choose damping or adapt the temperature so that basin escape is accelerated without making the dynamics unstable.
Useful7/10
Difficulty5/10
Novelty6/10
△ Mechanism confirmed, baseline not beaten
2026
Replace synchronous replicated-gradient computation with a bounded-staleness stream: at optimizer step t, aggregate one gradient for each data partition, using the newest completed evaluation even if it was computed at an earlier model version. Replicated partition placement makes the aggregate robust to stragglers, while pipelining ensures that each worker computes only one partition gradient per step.
Useful7/10
Difficulty6/10
Novelty6/10
✗ Failed on benchmark
2026
Replace a neural-network optimizer's globally fixed learning-rate geometry with an adaptive quadratic trust region. At every update, construct a local curvature model, accept or reject the step using the ratio between realized and predicted loss decrease, and expand or contract the radius accordingly; the same controller should automatically become conservative in nonconvex regions and Newton-like near a well-conditioned minimum.
Useful7/10
Difficulty6/10
Novelty6/10
△ Mechanism confirmed, baseline not beaten
2026
Partition optimizer state space into regions and assign each region a different update rule, such as two learning rates, momentum values, or preconditioners. Fit the local radial normal form of the resulting piecewise-smooth training dynamics and switch to the branch whose first nonzero coefficient predicts contraction toward the stationary point.
Useful7/10
Difficulty6/10
Novelty8/10
△ Mechanism confirmed, baseline not beaten
2026
Replace repeated iterations of an expensive high-dimensional update S with iterations of a lower-dimensional latent map T, then decode the resulting latent state with D. Train E, D, and T with explicit intertwining losses so that encoding a full update agrees with updating the latent state, and decoding a latent update agrees with applying the original update.
Useful7/10
Difficulty6/10
Novelty6/10
✗ Failed on benchmark
2026
Replace a stationary optimizer by a periodic two- or multi-phase schedule, such as alternating large and small learning rates, SGD and momentum, or gradients from different loss components. Stability is assessed over the complete period using the product of phase-wise linearized update maps, allowing a phase that is individually expansive to be safely combined with a contracting phase.
Useful7/10
Difficulty5/10
Novelty7/10
✗ Failed on benchmark
2026
Replace a capacity-penalty-only MoE router with a nonnegative shadow price for each expert, capacity bucket, or hardware resource. Route each token using predicted utility minus the relevant price, while computing a decomposed optimistic objective that certifies how much utility remains above the feasible routed value.
Useful7/10
Difficulty4/10
Novelty5/10
△ Mechanism confirmed, baseline not beaten
2026
Insert a finite-resolution observation channel between minibatch statistics and the optimizer update, then distinguish information that predicts useful future loss reduction from information that is present in the gradient but has no control value. Use the actionable representation to select the update and suppress increasingly fine, noisy measurements that do not improve progress.
Useful7/10
Difficulty6/10
Novelty7/10
✓✓ Beats tuned baseline
2026
Replace the usual projected gradient step with a relaxed projection in a positive-definite metric that changes with the current parameter state. The metric acts as a continuous preconditioner before projection, so updates can be large along poorly conditioned directions while remaining feasible.
Useful7/10
Difficulty5/10
Novelty6/10
△ Mechanism confirmed, baseline not beaten
2026
Replace a fixed temperature schedule in a population-based, derivative-free neural-network optimizer with a feedback controller driven by the entropy of candidate importance weights. When candidate losses are diffuse, the optimizer cools rapidly to exploit progress; when one or a few candidates dominate, cooling slows to prevent irreversible population collapse and loss of exploration.
Useful7/10
Difficulty5/10
Novelty6/10
✗ Failed on benchmark
2026
Use a two-mode optimizer: a learned preconditioned update for normal training and a bounded contractive fallback when the learned update is predicted to increase a monitored energy. Use separate entry and exit thresholds so minibatch noise does not cause rapid switching.
Useful7/10
Difficulty5/10
Novelty7/10
△ Mechanism confirmed, baseline not beaten
2026
Train a matrix-valued neural layer under an exact or near-exact Stiefel constraint while using an l1 or row-group sparsity penalty. During early training, use manifold proximal-gradient steps to identify a stable nonzero support; once the support stops changing, switch to Newton-CG steps restricted to the smooth intersection of the Stiefel tangent space and the fixed-support subspace. This can reduce the number of optimizer iterations needed to obtain sparse, well-conditioned projections.
Useful7/10
Difficulty6/10
Novelty7/10
✓✓ Beats tuned baseline
2026
Use the paper's lifetime law as a controller for training or rollout difficulty. Estimate the active perturbation bandwidth R of hidden states or forecast errors and reduce the residual gain, increase the dispersion order W, or inject controlled bandwidth whenever the estimated prethermal lifetime becomes too short.
Useful7/10
Difficulty5/10
Novelty8/10
✗ Failed on benchmark
2026
Replace the assumption that a minibatch gradient is fully Gaussian by a Gaussian center plus an explicit single-example big-jump correction. At each update, estimate the distribution of per-example gradient projections along the proposed update direction and use the predicted aggregate tail probability to reduce the step size or increase clipping only when the minibatch is in its non-Gaussian crossover regime.
Useful7/10
Difficulty5/10
Novelty7/10
△ Mechanism confirmed, baseline not beaten
2026
Use the extreme-eigenvector marginal test to decide whether a Kronecker preconditioner is condition-optimal, rather than blindly running expensive factor refinement. If the certificate fails, construct a low-cost factor correction from the mismatch between tensor marginals of the worst-conditioned spectral states and accept it only with a condition-number line search.
Useful7/10
Difficulty7/10
Novelty8/10
△ Mechanism confirmed, baseline not beaten
2026
Maintain a posterior over the effective stochastic-gradient noise scale and trigger expensive diagnostics or conservative optimizer changes only when uncertainty in that scale threatens a training-stability certificate. Unlike entropy-based exploration, the trigger depends on the predicted excess loss or stability gap caused by calibrating the optimizer to the wrong noise level.
Useful7/10
Difficulty5/10
Novelty7/10
△ Mechanism confirmed, baseline not beaten
2026
Replace ordinary low-precision multiply-add accumulation in selected neural-network reductions with a two-word floating-point accumulator updated by the paper's branch-free DW-FMA network. The high word retains the main sum and the low word stores the rounding residual, improving cancellation behavior without the control-flow divergence of conditional compensated summation.
Useful7/10
Difficulty5/10
Novelty6/10
△ Mechanism confirmed, baseline not beaten
2026
Build a low-dimensional neural-network geometry from trainable observables or probes instead of estimating the full Fisher matrix. Precondition the parameter gradient by the inverse variability of the probes and their parameter sensitivity, producing a task-adapted update that can remain usable for implicit models, heavy-tailed data, and parameter-dependent-support distributions.
Useful7/10
Difficulty6/10
Novelty6/10
✗ Failed on benchmark
2026
Regularize the local recurrent Jacobian by its spectral radius rather than imposing the overly conservative operator-norm condition $\|J\|_2<1$. This permits useful non-normal updates with transient amplification while explicitly pushing the asymptotic dynamics toward a stable fixed point.
Useful7/10
Difficulty7/10
Novelty7/10
✗ Mechanism failed
2026
Replace periodic client-to-server updates for an online neural-network head with event-triggered transmissions based only on local feature regressors and sufficient statistics, not on the current global parameter estimate. Each client transmits when its local Gram matrix or feature-response statistic changes enough that using the previously transmitted value would violate a prescribed perturbation bound. This should preserve exponential convergence in the strongly excited linear-head regime…
Useful7/10
Difficulty5/10
Novelty7/10
✗ Failed on benchmark
2026
Augment an optimizer with a measurable redistribution time for its internal state and compare it with the time scale of the changing gradient field. Use the resulting Damkohler number to interpolate between a fast quasi-static preconditioner and a history-preserving, non-equilibrium update, rather than applying one optimizer regime throughout training.
Useful7/10
Difficulty5/10
Novelty7/10