✗ Mechanism failed
2026
Replace a flat stochastic MoE router by a two-level continuous-time routing model: experts within a group communicate rapidly, while transitions between groups occur slowly. Use the effective class-level stationary distribution as a soft load-balancing prior, reducing routing oscillation while preserving expert specialization.
Useful6/10
Difficulty6/10
Novelty6/10
△ Mechanism confirmed, baseline not beaten
2026
Parameterize a relative-position or lag-decay function as a finite positive mixture of exponentials instead of learning arbitrary attention bias values. The resulting kernel is completely monotone on positive distances, so it is nonnegative, decreasing, and has alternating derivative signs; the mixture provides several learned memory scales without allowing oscillatory or unstable long-range biases.
Useful6/10
Difficulty4/10
Novelty7/10
✗ Mechanism failed
2026
Use a learned asymmetric Finsler-like cost instead of the symmetric Euclidean distance in attention logits. The metric has a Riemannian quadratic part and a directional drift term, while a differentiable barrier enforces the strong-convexity condition derived for the paper's extended $(\alpha,\beta)$-metrics. This lets each attention head prefer one direction in feature space without producing pathological, non-convex distance landscapes.
Useful6/10
Difficulty5/10
Novelty7/10
△ Mechanism confirmed, baseline not beaten
2026
Replace an independently generated discrete latent sequence or redundancy stream with a symmetric two-state Markov source whose transition probability is tuned or learned. Train the downstream transformer to reconstruct the semantic target after random insertions, while using the paper's insertion-capacity expression to select the latent rate and redundancy budget. The representation should preserve information under timing drift, repeated tokens, and inserted distractors better than iid token…
Useful6/10
Difficulty5/10
Novelty7/10
✗ Mechanism failed
2026
Bootstrap the optimizer curvature scale from a deliberately nondegenerate pair of gradient queries, then perform steepest descent in lp geometry with a local secant backtracking rule. The method does not require a supplied learning rate, smoothness constant L, initial distance R, or optimum value f*, and it automatically uses the dual norm associated with p.
Useful6/10
Difficulty5/10
Novelty6/10
✗ Mechanism failed
2026
Insert a weighted negative-semidefinite fourth-order mixing operator into a residual or state-space layer. Instead of learning an unconstrained token-mixing matrix, parameterize its dissipative component as Q = -a W^{-1} B^T W B, ensuring that this component cannot increase the chosen weighted feature energy. Use a boundary-aware finite-difference matrix B along the sequence axis, optionally with learnable banded coefficients while preserving the factorization.
Useful6/10
Difficulty5/10
Novelty5/10
△ Mechanism confirmed, baseline not beaten
2026
Use the weighted quadrature identity as a training or inference constraint for a compressed activation path: retain only a minimal set of binary evaluations and compute normalization or residual-energy statistics exactly on the modeled Rademacher component. This provides a zero-variance alternative to random activation subsampling for the represented subspace.
Useful6/10
Difficulty4/10
Novelty6/10
✓✓ Beats tuned baseline
2026
Estimate the temporal spectrum of each sequence channel using a locally private procedure, then apply a regularized inverse-square-root spectral filter before the sequence enters attention or an SSM. The filter removes predictable low-frequency or narrow-band redundancy while avoiding unstable amplification at frequencies where the private estimate is small.
Useful6/10
Difficulty5/10
Novelty7/10
✗ Mechanism failed
2026
Replace unconstrained combinations of several attention or adapter operations with a brace-style composition that inserts each operation into a distinct ordered interval of a base sequence. The resulting computation preserves the order of host and inserted operations and forbids crossing dependencies, producing hierarchical attention patterns with an explicit structural bias.
Useful6/10
Difficulty6/10
Novelty6/10
✗ Mechanism failed
2026
Represent candidate two-dimensional attention windows as dyadic rectangles and penalize local regions where many deeply embedded windows overlap. Use complementary horizontal and vertical depth exponents rather than independently penalizing one coordinate. The resulting router should reduce pathological concentration of sparse attention computation while preserving access to multiscale context.
Useful6/10
Difficulty6/10
Novelty7/10
△ Mechanism confirmed, baseline not beaten
2026
Construct a sparse attention support by solving multiple small perturbed assignment problems between query and key embeddings and taking the union of the selected optimal matchings. Use the resulting spanning tree as the only set of cross-token edges, with edge biases determined by empirical assignment frequency.
Useful6/10
Difficulty7/10
Novelty7/10
✗ Failed on benchmark
2026
Replace the usual hand-designed expert-load penalty with a heterogeneous survival penalty derived from a susceptibility distribution. Each expert receives an availability factor q_e=G(A_e), where A_e is its cumulative recent routing pressure and G_e is a learned or fixed mixture of exponentials; highly used experts are suppressed smoothly, while heterogeneous experts can have different resistance to pressure. The mixture produces adaptive curvature and long-tailed penalties that may reduce…
Useful6/10
Difficulty4/10
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
Represent an original neural-network block and a proposed rewritten block as constrained optimization formulations over inputs and trainable parameters, then certify that the rewrite preserves feasibility and the ordering of losses over a bounded domain. This gives a compiler or pruning pipeline a formal reject/accept gate instead of relying only on numerical regression tests.
Useful5/10
Difficulty7/10
Novelty8/10
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
Build a constrained autoregressive model whose initial logits are generated from a translation-invariant MPS associated with a local zero-mode construction. The MPS supplies a structured valid distribution before a Transformer residual is added, so the model starts on the constraint manifold instead of learning validity through a penalty.
Useful5/10
Difficulty6/10
Novelty6/10
Unverified
2026
Partition activations into dyadic magnitude bands and allocate sparse connectivity separately to heavy and diffuse coordinates. Protect high-magnitude coordinates with more reliable connections while using randomized flat connectivity for the many small coordinates, keeping the total number of nonzeros fixed.
Useful5/10
Difficulty6/10
Novelty7/10
Unverified
2026
Replace ordinary per-coordinate activation normalization or covariance whitening with a learned or frozen affine map based on a minimum-volume ellipsoid calibrated to the activation convex hull. Estimate the hull's Minkowski symmetry and use the bound \(\rho=\sqrt{d/\alpha}\) to set a principled clipping radius, where \(d\) is feature dimension and \(\alpha\) is the measured symmetry. This should reduce worst-case clipping of asymmetric activation distributions without requiring a much larger…
Useful5/10
Difficulty6/10
Novelty4/10
Unverified
2026
Replace raw powers or unconstrained polynomial spectral features with normalized Jacobi features whose amplitude is provably bounded on the entire input interval. Use trainable mixtures of these features in a positional encoding, graph spectral layer, or MLP front end, while preserving the theorem's normalization and optionally constraining the learned mixture norm.
Useful5/10
Difficulty4/10
Novelty7/10
Unverified
2026
Use the adjacency matrix of a vertex-transitive strongly regular graph as a fixed sparse attention or token-mixing mask. Every vertex has the same degree, and every pair of vertices has exactly one of two common-neighbor counts, giving predictable two-hop coverage and avoiding the degree and connectivity irregularities of random sparsification.
Useful5/10
Difficulty5/10
Novelty6/10
Unverified
2026
Replace random or learned routing hashes for tokens arranged on a d by d grid with a fixed family of mutually orthogonal anti-Latin squares. Each channel assigns exactly d of the d squared tokens to every bucket, while any two channels jointly distinguish every grid position. The resulting router has deterministic load balance and multi-view positional diversity.
Useful5/10
Difficulty5/10
Novelty7/10
Unverified
2026
Replace a dense token- or channel-mixing matrix with a product of local braid generators acting on adjacent coordinates. Each generator is an exactly invertible 2-by-2 transformation, while the braid and far-commutativity identities give multiple equivalent factorizations of the same global operator. This creates a sparse, reversible mixer with O(kn) cost for a braid word of length k, rather than O(n^2) cost for a dense matrix.
Useful5/10
Difficulty4/10
Novelty7/10
Unverified
2026
Construct positional or relative-position features as a nonnegative mixture of lattice cosine functions instead of independently signed sinusoidal features. The resulting bias is the Fourier transform of a positive discrete measure with explicitly bounded spectral support, while the mesh and degree can be initialized in the paper's dense-but-controlled frequency regime.
Useful5/10
Difficulty4/10
Novelty7/10
Unverified
2026
Use the paper's multicycle result to distinguish useful parameter motion from internally circulating optimizer activity. Add an auxiliary two-cycle diagnostic to an optimizer or recurrent training loop: one cycle represents net loss-improving motion, while another represents momentum or noise circulation that can remain active even when the net parameter update is nearly zero. Penalize or throttle this hidden circulation to prevent apparent convergence from masking high update variance and…
Useful5/10
Difficulty6/10
Novelty7/10