Every idea extracted from recent arXiv mathematics papers — verified and unverified. Click an idea to open its full card; badges show the empirical verdict.
Replace dense token attention with K asynchronous local nonlinear consensus updates on a sparse token graph. Each selected token is moved to the unique incident p-mean of its neighbors, producing a tunable message-passing operator that is averaging for p=2 and nonlinear for other p.
Construct a sparse directed attention mask over ordered tokens by treating token pairs as edges labeled by time or position. Apply a recursive three-hop spanner construction so that every dense temporal route has a sparse route using at most three times as many hops. Run attention only on the selected edges and use three sparse message-passing layers, or a small learned hop-composition block, to approximate dense causal attention.
Represent an intermediate hidden-state matrix as a low-rank factorization and replace a dense residual update by a repeated contractive step followed by hard SVD truncation. Increase or decrease the rank automatically so that the discarded Frobenius-norm tail is below a specified tolerance, providing an explicit accuracy-memory tradeoff instead of fixing the rank in advance.
Construct each hierarchy transition as a coarse-graining map together with a local decoder that recovers the retained local representation. This creates a reversible-information path for features needed by later computation while still allowing high-frequency or locally irrelevant components to be discarded.
Add a recurrent associative matrix to each selected transformer layer so recent key-value relationships can be retrieved without retaining every past token or performing gradient updates. The matrix uses input-dependent retention and write gates, but retrieval is always performed from the pre-write state, preventing the current target from leaking into its own prediction. Frobenius-norm clipping makes the recurrent memory bounded and provides a direct stability control.
Replace an unconstrained residual block by a first-order gradient-flow correction whose energy contains first-, second-, and third-difference penalties, mirroring the paper's higher-gradient gravitational energy. The correction suppresses high-frequency modes while retaining a trainable nonlinear residual branch, and its step size can be chosen from an explicit spectral stability bound.
Replace an ordinary local convolution or token-mixing block by two nonnegative feature populations A and B that diffuse and drift along the spatial or token axis, with transport rates increasing quadratically with local population and with an optional directional bias. Add a local A plus B to empty reaction so mutually conflicting feature mass is removed rather than merely averaged. The module should produce adaptive competition, and its isolated relaxation should exhibit a measurable t raised…
Replace the dense output of selected linear projections with a two-sided magnitude threshold that emits zero for small values but preserves signed large values. Learn one positive threshold per projection, or optionally one threshold per output channel, so the network discovers where sparse events can be removed while retaining outlier information.
Impose a screened pair-correlation dynamics on stochastic neural replicas so that correlation fluctuations relax locally instead of propagating across the entire representation. The key control knob is a learned or scheduled relaxation rate \(\mu_{FB}\), which predicts a measurable correlation length \(\xi_{FB}=\sqrt{D_{eff}/\mu_{FB}}\). This can be used as a locality regularizer for token representations, diffusion trajectories, or recurrent hidden states.
Use cumulative control measures as the input to a neural surrogate instead of the full sequence of control values. For a quasi-static free-boundary system satisfying the paper's average-pressure path-independence assumption, two nonnegative control histories with identical integrals up to time t should produce the same state at t, allowing a smaller training input and fewer distinct control trajectories in the dataset.
Replace the linear phase progression in a positional encoding or rotary attention mechanism with a deterministic quadratic phase. The resulting position signal is generated by an irrational rotation with linearly changing increments, and the paper proves that its infinite diffraction measure is purely absolutely continuous, suggesting disorder-like spectral coverage without random sampling.
Partition neural modules into two empirically identified reliability or noise classes and restrict their communication graph to a two-block stochastic block model. Allocate a fixed connectivity budget across within-class and cross-class edges using a water-filling update that favors block pairs producing the largest increase in validation utility. The resulting layer is sparse and modular, with a testable prediction that optimal connectivity concentrates on a few block pairs rather than…
Replace fixed LoRA factors with a rank-adaptive moving subspace whose columns are augmented using derivative information from several Runge–Kutta stages. The optimizer integrates a matrix-valued gradient-flow approximation inside this enlarged left/right basis, allowing high-order motion of the adapter subspace while retaining a low-rank parameterization.
Add a local Jacobian spectral regularizer and an initialization sweep to steer a looped transformer away from uncontrolled near-unit dynamics. The goal is to prevent examples from entering a fold-critical regime with very long relaxation times, or alternatively to deliberately target a controlled critical regime when adaptive test-time compute is useful.
Add a small dynamical state on the transformer module graph and use it to control adaptive computation, but reject controller parameters whose discrete-time update has latent roots outside the unit disk. The state can modulate halting thresholds, residual-block gains, and memory gates; the certificate applies to the controller integrator and prevents unstable oscillations or exploding internal control signals during long adaptive-depth rollouts.
Construct a residual neural network or graph message-passing layer whose skip edges are sampled with probability proportional to their distance as $|i-j|^{-(1+\sigma)}$, while retaining a small local backbone. The paper's mechanism predicts that coarse-grained propagation is governed by the long-range kinetic operator and is therefore asymptotically insensitive to the particular Bernoulli graph realization, yielding controllable superdiffusive information transport without dense all-to-all…
Construct a sparse fixed orthogonal mixer by repeatedly applying pi/4 rotations to randomly matched pairs of feature coordinates. Place this mixer before top-k feature pruning, sparse projection, or activation quantization so that information is spread across coordinates without using a dense random matrix.
Replace dense token-to-token attention in a controlled ablation with a cyclic order-a recurrence whose coefficients are periodic across positions. The resulting mixer has linear cost in sequence length for fixed recurrence order and can enforce a fixed signed periodic boundary condition, providing a compact structured alternative to local attention or a lightweight state-space model.
Treat a small activation, gradient, or parameter perturbation as a seed and measure the number of newly affected downstream units or layers. Use the estimated branching ratio to control the optimizer step size or residual gains, keeping training in a subcritical regime where perturbation cascades have finite expected size instead of amplifying through the whole network.
Replace unconstrained or entropy-regularized MoE routing with a minimally disruptive update that preserves a lower bound on the log-determinant of the experts' weighted output span. The router still tracks the desired mixture, but a projection prevents the active experts from becoming linearly redundant or collapsing onto a low-rank subset.
Treat a selected neural submodule as an open dynamical system embedded in the rest of the network. Regularize it to contain internal modes that are simultaneously reachable from many external features and observable through many external outputs, rather than behaving as a one-sided receiver, broadcaster, or disconnected read/write split.
Replace a global attention truncation rule with a per-query local-tail budget. For each query token, retain nearby or high-priority keys until the estimated discarded interaction strength is below a target epsilon; this uses the paper's central distinction between local tail mass and the extensive norm of the discarded operator. The resulting attention pattern can allocate long-range computation only to tokens whose local tail is large.
Replace a single local message-passing or convolution operator by a spectrally controlled mixture of fractional and ordinary diffusion. The exponent σ is learned or scheduled, while a crossover gate forces the model to change parameterization near the renormalization-group threshold σ*=2, allowing long-range propagation when useful without retaining an unnecessarily nonlocal operator at short scales.
Replace a dense neural interaction graph by a dynamically activated graph whose edge $(u,v)$ is retained only when its effective coupling exceeds the local spacing of response modes. The network remains sparse below the connectivity transition but becomes globally communicating once a giant component forms, providing a controllable alternative to arbitrary magnitude pruning.