Architecture ideas

Research ideas extracted from mathematics papers, categorized as Architecture.

Unverified 2026

Diffeomorphic gauge-fixing layer

Insert a differentiable spatial canonicalization module before a neural dynamics model. It estimates a smooth invertible coordinate transformation that places each input field in a common gauge relative to a reference template, predicts the next state in that gauge, and maps predictions back to the original coordinates. The module should reduce the need for the dynamics network to relearn identical laws under many smooth spatial reparameterizations.

Useful6/10
Difficulty6/10
Novelty6/10
Paper: The Right Space for Dynamics: Numerics with Diffeomorphism Equivariance arXiv:2607.06536
Unverified 2026

Periodic CMV Unitary Recurrent Layer

Replace a dense recurrent transition matrix with a periodic CMV-style product of alternating local 2x2 unitary cores. The transition is exactly norm-preserving, has O(n) trainable parameters under periodic tying, and can be applied through local factor operations rather than stored as an n-by-n matrix. Use turnover refactorization when changing the ordering or boundary connection of cores, enabling a compact cyclic unitary state-space layer.

Useful6/10
Difficulty5/10
Novelty5/10
Paper: Fast computation of eigenvalues of periodic CMV matrices arXiv:2607.06400
Unverified 2026

Fractional Mahalanobis radial head

Replace a binary classifier's unconstrained final logit with a differentiable likelihood-ratio head based on two squared Mahalanobis radii in a learned embedding space. Approximate the shared radial generator with a small fractional-power basis, allowing the head to model heavy-tailed class geometry that an affine QDA logit cannot represent while remaining much smaller than a generic nonlinear head.

Useful6/10
Difficulty5/10
Novelty6/10
Paper: Closed-form fractional radial links for elliptical Mahalanobis discriminant analysis arXiv:2607.06089
Unverified 2026

Lipschitz-Free Metric Pooling

Replace coordinate-wise mean pooling of metric-valued items with a finite representation of their free integral. Each item x in a pointed metric space M is represented through evaluations of learned Lipschitz probes, and the pooled feature is the weighted integral of those probe values. A dual Lipschitz critic estimates the free-space norm of differences between pooled groups, making the representation sensitive to metric geometry while remaining permutation-invariant.

Useful6/10
Difficulty5/10
Novelty7/10
Paper: Analytic integration of metric-valued functions in Lipschitz free spaces arXiv:2607.06049
Unverified 2026

Support-Budgeted Hamming Polynomial Layer

Replace the first dense layer on q-ary categorical features by a Fourier interaction layer containing only monomials whose coordinate support is at most s. Use a Bohnenblust–Hille-inspired quasi-norm on coefficients, separately for each interaction order, to prevent a small number of high-order interactions from dominating the output. The resulting model has an explicit interaction-order knob and can be tested against a dense MLP at matched parameter count.

Useful6/10
Difficulty5/10
Novelty7/10
Paper: Support-Sensitive Bohnenblust-Hille Inequalities and Local Invariants on Hamming Schemes arXiv:2607.05594
Unverified 2026

Bernstein resolvent activation

Replace an unconstrained scalar activation or nonnegative gate with a finite positive mixture of rational Bernstein basis functions. The learned function is monotone and concave on the nonnegative half-line, while its derivatives have controlled alternating signs; this can prevent pathological feature amplification and gives an interpretable shape prior. Use the paper's sharp exponent restriction τ≤1/2 rather than treating the power as an arbitrary hyperparameter.

Useful6/10
Difficulty4/10
Novelty7/10
Paper: Riccati Reductions for Modified Bessel Ratios: Bernstein Positivity, Exact Certificates, and Transfer Obstructions arXiv:2607.05538
Unverified 2026

Heisenberg latent upsampler

Represent each latent state as a Heisenberg-group element and replace Euclidean interpolation in an upsampling or recurrent transition block by a four-point horizontal refinement plus the exact central signed-area correction. The module preserves the geometry of noncommutative composition, allowing the central latent coordinate to encode path-dependent information that ordinary coordinate-wise interpolation discards.

Useful6/10
Difficulty5/10
Novelty8/10
Paper: A Heisenberg Subdivision Scheme with Central Smoothness Loss arXiv:2607.05446
Unverified 2026

Dual-unitary recurrent state block

Replace a generic recurrent transition with two coupled unitary transitions that share one block column and differ by a sign on the other block column. Each transition preserves hidden-state norm exactly, while the structured difference gives a controlled two-path recurrent architecture for long-context modeling.

Useful6/10
Difficulty5/10
Novelty5/10
Paper: Operator model and a trace formula for pairs of unitary operators arXiv:2607.05334
Unverified 2026

Gain-Rigid Sparse Attention

Construct a sparse attention or message-passing graph as a gain graph, where every directed edge carries a relative frame label and reverse edges carry the inverse label. Grow the graph using the paper's 2-extension operation: replace two old edges by a new vertex connected to their four endpoints, while preserving the relative gain products. The resulting mask is intended to preserve global information flow under controlled sparsity and to avoid isolated components and brittle bridges commonly…

Useful6/10
Difficulty6/10
Novelty7/10
Paper: Rigidity on compact surfaces through hyperbolic symmetries arXiv:2607.05023
Unverified 2026

Pole-Certified SSM Initialization

Extract a small set of stable exponential modes from an observed neural sequence and use them to initialize a diagonal or block-diagonal state-space model. Hankel-pencil eigenvalues propose the modes, while persistence across shifts and contour margins reject modes caused by noise or a short-lived background.

Useful6/10
Difficulty5/10
Novelty7/10
Paper: Determinant Characteristics and Argument-Principle Certification for Visible Poles in Meromorphic Continuation arXiv:2607.04568
Unverified 2026

Kolmogorov-Lie Unitary Layer

Build an input-conditioned unitary transformation as an ordered product of exponentials of anti-Hermitian matrices, with each factor controlled by a univariate function of one input coordinate or one learned scalar projection. This replaces a dense multivariate matrix-valued controller with separable scalar nonlinearities while preserving exact unitarity at every forward pass.

Useful6/10
Difficulty6/10
Novelty7/10
Paper: Quantum Kolmogorov--Arnold representation theorem for continuous unitary-valued maps arXiv:2607.03187
Mechanism works 2026

Degree-Corrected Hierarchical Router

Replace a flat MoE or graph-pooling assignment with recursive partitions selected by interaction evidence after removing each item’s expected degree effect. Tokens, nodes, or examples that are frequently active for purely popularity-related reasons should not automatically form an expert or cluster. Recursion stops when a candidate split has nonpositive degree-corrected evidence, producing an adaptive hierarchy rather than a fixed number of equally sized groups.

Useful6/10
Difficulty6/10
Novelty7/10
Paper: Community structure of the pseudofractal web arXiv:2607.03010
Unverified 2026

Fractional Boundary-Factored Neural Solver

For a fractional Dirichlet problem, replace a free coordinate network N_theta(x) with u_theta(x)=d(x)^a N_theta(x), where d(x)=dist(x,boundary) and 0<a<1 is the fractional order. Train the regular quotient v_theta=u_theta/d^a=N_theta and use a weighted gradient loss that reflects the paper's boundary estimate.

Useful6/10
Difficulty4/10
Novelty7/10
Paper: The structure of solution spaces for fractional-order operators, with gradient estimates arXiv:2607.02312
Unverified 2026

Fourier-Calibrated Nonlocal Feature Gradient

Augment a CNN with a nonlocal feature-gradient branch that compares each feature vector with a kernel-weighted neighborhood rather than using only pointwise or local convolutional interactions. Regularize this branch using the paper's Fourier multiplier energy, which penalizes feature oscillations according to the kernel spectrum and approaches an ordinary local-gradient operator as the interaction radius tends to zero.

Useful6/10
Difficulty4/10
Novelty6/10
Paper: A Variational Nonlocal Phase-Field Model for Dynamic Fracture in Elastic Solids arXiv:2607.01881
Unverified 2026

Totally-positive bidiagonal mixer

Replace a dense channel or token-mixing matrix with a product of positive bidiagonal factors, so information propagates through a controlled sequence of local couplings rather than arbitrary signed interactions. Initialize the factors from the paper's barycentric-subdivision factorization, then learn positive diagonal and off-diagonal parameters; the resulting map is structured, parameter-efficient, and constrained to remain totally positive.

Useful6/10
Difficulty5/10
Novelty8/10
Paper: Total positivity of transformation matrices for uniform subdivisions arXiv:2607.01577
Unverified 2026

Agnostic Geometry-Prior Mixer

Train an unconstrained branch and a geometry-aware branch in parallel, then learn how much to trust the analytic branch. This preserves the benefit of explicit geometry on correctly specified tasks while allowing the model to ignore a misleading or irrelevant prior.

Useful6/10
Difficulty3/10
Novelty7/10
Paper: Geometry-Aware R-Structured Kolmogorov-Arnold Networks arXiv:2607.01449
Unverified 2026

Monotone Singular-Value ICNN Envelope

Replace a generic neural constitutive law or energy model with an ICNN that consumes the positive singular values of a deformation-like matrix and is convex and coordinatewise nondecreasing in those inputs. Train it as a lower approximation to a nonconvex target energy, so the network acts as a computationally cheap sufficient polyconvex-envelope surrogate rather than merely interpolating unstable samples.

Useful6/10
Difficulty4/10
Novelty4/10
Paper: Compression of Polyconvex Envelopes of Isotropic Functions via Monotonic Input Convex Neural Networks arXiv:2607.01055
Unverified 2026

Tropical support-restricted PINN

Represent the PINN solution in a restricted polynomial or Taylor basis whose exponent set is supplied by tropical support analysis, instead of asking an MLP to discover the local series structure from scratch. The restriction removes coefficients that cannot occur in the formal solution, reducing trainable degrees of freedom and preventing spurious low-order or singular terms.

Useful6/10
Difficulty5/10
Novelty8/10
Paper: Tropical Geometry as a Restricted Architecture for Physics-Informed Neural Networks: Applications in Nonlinear Fluid-Structure Examples arXiv:2607.00237
Unverified 2026

Pole-safe rational neural layer

Replace an unconstrained feature vector entering a rational or resolvent-like neural operator by a polynomial feature whose first nonzero Taylor coefficient lies in a pole-safe subspace. For a pole of order m, the simplest guaranteed construction is psi(z)=(z-beta)^m v, which makes Q(z)psi(z) bounded even when Q(z) diverges. For lower-order cancellation, solve linear constraints among Taylor coefficients of psi so that all negative Laurent powers vanish.

Useful6/10
Difficulty5/10
Novelty8/10
Paper: Construction of Pole Cancellation Functions at Ordinary Poles of Operator-Valued Functions arXiv:2607.00097
Unverified 2026

GQL Safe Residual Layer

Insert a scalar flux-correction-style limiter after a neural operator predicts a conservative state or residual. Interpolate between a known-admissible baseline state and the learned high-order candidate, choosing the largest coefficient that satisfies a geometric family of linear inequalities encoding positive density, positive pressure, and subluminal velocity. This retains as much of the neural prediction as possible instead of independently clipping physical variables.

Useful6/10
Difficulty5/10
Novelty7/10
Paper: GQL-Based Physical-Constraint-Preserving High-Order Finite Difference Schemes for Special Relativistic Hydrodynamics in Arbitrary Dimensions arXiv:2606.31992
Unverified 2026

Positive Mellin Mixture Gate

Replace a free-form order-dependent gate with a positive mixture of Mellin powers $(1+s)^{-a}$. This gives a small, interpretable module whose response across the order variable is automatically generated by a positive measure and therefore inherits complete monotonicity, log-convexity, and Hankel-moment structure.

Useful6/10
Difficulty4/10
Novelty8/10
Paper: Order-Moment Transport and Hankel Determinants in Special-Function Inequalities arXiv:2606.31647
Unverified 2026

Fast-Class MoE Router

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
Paper: The Invariant Measure of Multiscale Markov Chains via Fast Arborescence Factorization arXiv:2606.31596
Unverified 2026

Completely Monotone Multiscale Attention Decay

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
Paper: Bernstein Functions at Work: Coalescents, Copulas, and Subordination arXiv:2607.04467
Unverified 2026

Markov-coded synchronization-robust latent streams

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
Paper: The Insertion List-Decoding Capacity and an Improved Bound on the Deletion List-Decoding Capacity arXiv:2607.03989