Attention variants, state-space and recurrent cells, normalization and token-mixing schemes — each tested against the standard block it replaces.
Augment a neural policy with deterministic DFA states for the task objective and safety constraint, then select among objective-specific policy heads using those states. Before either target is reached, execute a mixed policy; after one target is reached, switch permanently to the policy specialized for the remaining target.
Insert a differentiable equilibrium layer between a neural payoff/state encoder and the final action recommendations. The layer parameterizes a joint recommendation object and enforces all unilateral-deviation inequalities as positive-semidefinite constraints, preventing the network from producing recommendations that agents have a strict incentive to disobey. A quantum-inspired density-matrix parameterization can model correlated recommendations using PSD matrices rather than factorized action…
Replace an unconstrained learnable distance-bias function in a graph neural network or distance-aware attention layer by a Bernstein approximation of a positive-definite circular kernel. The resulting kernel is a degree-n polynomial in normalized distance while preserving positive semidefiniteness of every finite Gram matrix on the circle, preventing training from producing an invalid covariance-like similarity structure.
For local structures with a forward/reverse ambiguity, expose both ordered directions and add one explicit orientation bit. This creates a shared bidirectional positional encoder that can distinguish reflected neighborhoods without maintaining two completely independent directional encoders.
Replace exhaustive optimization of N binary gates by the geometrically admissible sign patterns induced by projections onto a common direction. For two-dimensional gate vectors, enumerate angular cells exactly; for higher-dimensional vectors, sample directions and evaluate only the induced configurations.
Replace repeated fine-scale message passing on a recursively constructed graph by a finite-state compositional module that summarizes each cell through boundary connectivity states. Child summaries are combined only through admissible configurations, so the network can represent globally consistent paths and forests without enumerating full walks. The same update map is shared across all hierarchy levels, giving logarithmic-depth computation in the graph diameter and a strong inductive bias for…
Represent a neural-network weight tensor by rank-one terms whose mode factors are selected from shared orthonormal bases, and impose the same basis alignment across tensor flattenings. During or after training, retain the largest coefficients to obtain a structured truncation analogous to truncated SVD. This should produce better-conditioned tensorized layers than unconstrained CP factors while preserving a directly controllable accuracy/compute tradeoff.
Construct a 2D recurrent or residual neural lattice with slowly varying local couplings, while parameterizing those couplings so that an anisotropy invariant remains constant across all spatial and depth locations. The network obtains controlled local propagation velocities rather than arbitrary inhomogeneous amplification, enabling depth-dependent receptive fields while preserving near-critical signal propagation.
Construct a neural layer with two independently ordered transformations and train its operators to satisfy the paper's diamond equations, so that applying direction 1 then direction 2 gives the same result as direction 2 then direction 1. Unlike ordinary weight sharing, the mixed identity permits noncommuting operators whose interaction defects cancel exactly.
Build a recurrent ReLU module that evaluates a binary refinement cascade using a fixed-dimensional state and shared cell weights. Replace hard binary digit selection with two overlapping circle coordinates; switch between their affine state updates only at points where the two candidate updates agree, so the switch is an exact continuous piecewise-linear ReLU operation rather than a multiplicative gate.
Project each FFN residual update onto the tangent space of the current token residual direction before adding it to the stream. This preserves the component that changes representation direction while suppressing norm-only motion, which may reduce residual-norm drift and aggregation-induced representation collapse.
Equip a recurrent or state-space layer with multiple noncommuting transition operators and regularize the span of finite operator words applied to the input injection matrix. This discourages hidden directions that cannot be reached from the input and may improve long-range input influence, gradient propagation, and robustness under operator switching.
Replace a conventional signed-graph message-passing layer with two coupled feature channels: a fixed channel invariant under switching and an anti-fixed channel that changes sign under switching. Unsigned aggregation updates invariant features, while signed aggregation updates anti-invariant features, implementing the paper's sphere-plus-involution representation at the hidden-state level.
Construct an unrolled phase-retrieval network that begins with an isotropic Gaussian estimate rather than a spectral initializer. Retain the AMP residual correction and Onsager subtraction, but learn the scalar measurement denoisers and step sizes; use several random starts and select the iterate with the lowest measurement residual.
Replace a learned sequence-mixing matrix with a structured lower-triangular Sprugnoli operator whose square is exactly the identity. Applying the same operator in reverse reconstructs activations exactly, so it can be used as a reversible Transformer mixer or reversible channel permutation while retaining nontrivial long-range mixing.
Replace raw pairwise attention or graph-edge scores by exact U-centered residuals, removing additive effects attributable to either endpoint. The resulting scores represent interaction beyond independent source and destination biases and satisfy zero row sums, preventing a few high-degree or high-activation tokens from dominating relational aggregation.
Replace a recursive product implementation of a rational spectral filter with an additive sum of independently evaluated resolvents. Use the layer on a graph Laplacian, token-similarity operator, or other sparse feature operator to obtain a high-order filter without multiplicative roundoff and gradient amplification; the independent solves can also be batched or distributed across devices.
Store a finite library of successful robot configurations or action-conditioned waypoints and construct a smooth soft minimum of their distances. Use the negative distance gradient as a structured action prior, add a learned residual policy, and pass the combined action through a quadratic-program safety layer. This gives a neural controller an explicit attraction basin toward demonstrated solutions while preventing violations of known state constraints.
Replace dense attention or strictly local windows with a refreshed sparse communication graph whose edge lengths follow a truncated power law. For sequence positions or image patches at distance $r$, sample edges with probability proportional to $(1+r)^{-(d+\alpha)}$; in the regime $0<\alpha<d$, repeated layers should spread information across the sequence much faster than a local window while using only $O(nk)$ edges per layer. The main test is whether this preserves long-range accuracy at…
Replace unconstrained low-rank compression of a neural state with an augmented basis that always contains vectors representing known conserved quantities or diagnostically important linear statistics. After each learned transition, project the state back onto the affine constraint set with an exact minimum-norm correction, preventing rank truncation and model error from accumulating in those statistics.
Attach a symbolic sparse head to a neural encoder instead of using a dense final MLP. The head evaluates a library of learnable power-law and interaction terms on nonnegative learned features, jointly optimizes linear coefficients and exponents, and removes inactive terms with coefficient sparsity. This should provide a compact model with better relative-error behavior on positive targets spanning several orders of magnitude.
Replace or augment a mixture-of-experts router with a relative transverse-curvature score computed between experts, rather than relying only on the router MLP logits. Experts that provide a broader, less stiff local response in task-relevant directions receive higher routing probability, while common nuisance or spectator directions cancel from the comparison. The score is invariant under a common linear reparameterization of the routing coordinates and can be restricted to a low-dimensional…
Replace a conventional graph message-passing layer with a finite-horizon stochastic propagation process containing susceptible, infected, and removed feature states. Messages spread along active infected-to-susceptible edges, while infected nodes are simultaneously deleted at a rate proportional to their susceptible-neighbor count. This provides explicit propagation control and anti-oversmoothing dynamics instead of repeatedly averaging over every neighbor.
Augment standard Transformer positional embeddings with coordinates generated by the paper's signed-base digit expansion. Previous binary digits determine the sign and scale of later contributions, while a two-state Markov chain controls correlations between digits. This supplies multiscale positional structure using a small number of transition and base parameters.