Architecture ideas

Research ideas extracted from mathematics papers, categorized as Architecture.

Mechanism failed 2026

Sparse symbolic vector-field layer

Replace an opaque MLP vector field with a stack of trainable symbolic primitives that can express linear terms, monomials, products, and related analytic operations. Apply an L1 penalty and prune small primitive coefficients after rollout training, yielding a compact dynamics module that is cheaper to evaluate and easier to inspect.

Useful7/10
Difficulty6/10
Novelty6/10
Paper: Symbolic Neural ODEs: Learning interpretable models from time-series data arXiv:2608.22112
Mechanism failed 2026

Delay-Ring Replay Memory

Replace a directed sequence-memory chain with a circular recurrent state propagated by a learned delayed convolution. The same learned kernel can support forward and reverse replay because replay direction is a dynamical mode of the ring, rather than requiring plasticity to explicitly learn both forward and backward synapses.

Useful7/10
Difficulty5/10
Novelty8/10
Paper: Forward and reverse delay-driven hippocampal replay without symmetric plasticity arXiv:2608.21814
Unverified 2026

Pressure-Controlled Neural IFS

Construct a generative or recurrent neural architecture with several contractive or mildly expanding branches, and explicitly control the geometric complexity of its invariant set using the sub-additive singular-value pressure of branch-Jacobian products. Instead of regularizing only the operator norm, the model can preserve anisotropic directions while targeting a desired attractor dimension, potentially improving coverage of structured data without uncontrolled folding or collapse.

Useful6/10
Difficulty6/10
Novelty8/10
Paper: Dimensions of surface repellers and attractors of non-linear planar IFSs arXiv:2608.30744
Unverified 2026

Trainable two-sided activation gate

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.

Useful6/10
Difficulty4/10
Novelty5/10
Paper: Event-Driven Language Models with Sparse Neural Activity for Neuromorphic Hardware arXiv:2608.30439
Unverified 2026

PCA-Discovered Implicit Latent Dynamics

Replace an explicit recurrent transition with a learned descriptor relation in latent space, allowing some latent coordinates to satisfy algebraic constraints rather than being numerically integrated. Fit the relation using total-least-squares or iterative PCA on the jointly observed trajectory, so noise in every channel is treated symmetrically and the model can discover whether the latent system is index-0 or index-1.

Useful6/10
Difficulty6/10
Novelty7/10
Paper: Subspace Based Identification of Errors-in-Variables Linear Descriptor Systems arXiv:2608.30259
Unverified 2026

Composite-cycle interface search

Use exhaustive finite-state analysis as an architecture-search objective for coupled recurrent motifs. Instead of independently tuning a normalization gate and a WTA gate, enumerate their possible interfaces and select couplings that create group action only on a small joint image set, yielding controlled reversible subdynamics embedded in an otherwise dissipative system.

Useful6/10
Difficulty6/10
Novelty9/10
Paper: "More Is Different'' in Neural Circuits: Algebraic Emergence of Effective Theories in Canonical Recurrent Motifs of Biological Neuronal Networks arXiv:2608.30231
Unverified 2026

Reciprocal Paired-Density Network

Represent a neural density or feature field by two positive reciprocal branches whose product is the modeled density, analogous to the forward and backward fields in the paper. Add stochastic branching perturbations to the two branches and train their cross-covariance so that the diagonal paired density matches the target while off-diagonal correlations remain finite-range. This creates a structured alternative to an unconstrained single-field uncertainty representation.

Useful6/10
Difficulty6/10
Novelty7/10
Paper: Branching stochastic mechanics. I. Clustering and connected correlations within a branching-process representation of the Schrödinger equation arXiv:2608.29807
Unverified 2026

Mass-Conserving Puncta Router

Insert a differentiable reaction-diffusion layer that converts dense token or pixel features into sparse, spatially coherent routing masks. Two competing orientations form complexes through conserved monomer reservoirs, so local assignments can cluster while opposite assignments mutually exclude one another instead of independently activating at the same location. The layer can be used as a soft-to-hard MoE router, attention-mask generator, or object-part grouping module.

Useful6/10
Difficulty5/10
Novelty7/10
Paper: Clustering versus sorting: a mass-conserving reaction-diffusion model of planar polarity puncta arXiv:2608.29679
Unverified 2026

Cumulative-Control Surrogate Encoder

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.

Useful6/10
Difficulty4/10
Novelty8/10
Paper: Online Gate-Driven Flow Control in Resin Transfer Moulding Using a Neural-Network Surrogate arXiv:2608.29521
Unverified 2026

Relative-Degree-Gated Passive Neural State Space

Construct a neural state-space model with an explicit first-order input-to-output path instead of forcing every output to depend only on deeply propagated hidden states. Penalize or reject learned linearizations whose transfer matrix has relative degree greater than one, then train a storage-function certificate for the remaining passive dynamics. This preserves the paper's relative-degree compatibility condition while allowing high-order internal memory.

Useful6/10
Difficulty6/10
Novelty7/10
Paper: Relative-Degree Wall Restricts Passivity-Based Stability Analysis in Inverter-Dominant Grids arXiv:2608.29474
Unverified 2026

Hard-Saturated Neural Feedback

Build actuator or parameter constraints directly into the neural controller using a differentiable hard-saturation map rather than penalizing violations after the fact. This makes the Lyapunov certificate apply to the actual bounded controller and prevents training from exploiting unrealistically large actions.

Useful6/10
Difficulty3/10
Novelty4/10
Paper: Learning neural controllers for nonlinear systems from data arXiv:2608.29303
Unverified 2026

Stability-calibrated Sinkhorn attention

Replace independently normalized attention or routing weights with an entropic doubly stochastic transport plan, while choosing its regularization ε using the paper's explicit statistical-stability bound. Increase ε when residual inversion or minibatch fluctuations are amplified, and decrease it only when the estimated bound permits sharper assignments.

Useful6/10
Difficulty6/10
Novelty6/10
Paper: Uniform Statistical Convergence of Empirical Sinkhorn Potentials with Exponential and Polynomial Dependence on the Regularization Parameter arXiv:2608.29152
Unverified 2026

Laguerre-Optimal Positive Delay Filter

Replace an Erlang delay or exponential smoothing cascade in a recurrent or state-space layer by a positive rational kernel of the form \(\kappa(u)=C e^{-a u}p(u)^2\). Choose the degree-\(m\) polynomial by deleting the adjacent pair of Laguerre zeros with smallest relative gap from \(L_{m+2}\), then rescale the resulting density to unit mean. This preserves a nonnegative impulse response while reducing temporal jitter relative to Erlang filters.

Useful6/10
Difficulty6/10
Novelty7/10
Paper: Least Variability in a Polynomial-Square Class of Rational Kernels arXiv:2608.29143
Unverified 2026

Quadratic-Chirp Positional Rotation

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.

Useful6/10
Difficulty4/10
Novelty6/10
Paper: Pseudorandomness and Diffraction arXiv:2608.28917
Unverified 2026

Spectral Sign-Balanced Update Blocks

Represent a block of candidate neural updates or adapter components by symmetric influence matrices and select one sign for each component so their aggregate spectral effect is small. This imports matrix discrepancy into low-rank adapters, expert aggregation, or structured quantization, where controlling the worst direction of interference may be more useful than minimizing entrywise error.

Useful6/10
Difficulty5/10
Novelty7/10
Paper: A Proof of the Matrix Spencer Conjecture arXiv:2608.28816
Unverified 2026

Stable Hyperplane Feature Layer

Replace a dense directional feature map with a finite stable-zonotope layer: a small set of large random vectors creates a hyperplane arrangement over the input sphere, while a cheaper small-jump residual gives continuous features inside each cell. The layer exposes an explicit compute-versus-resolution knob through the jump cutoff and can be used either as an embedding before an MLP or as a lightweight router for mixture-of-experts blocks.

Useful6/10
Difficulty5/10
Novelty7/10
Paper: Hausdorff Dimension of the Set of Extreme Points of a Random Countable Stable Zonotope arXiv:2608.28004
Unverified 2026

Minor-preserving complex network layer

Parameterize a complex linear layer as a product of sparse triangular network factors whose positive modulus version is totally nonnegative. The layer can use phase cancellation for expressive transformations, while selected minors remain bounded by explicitly computable positive minors, giving a structured alternative to unconstrained dense complex weights.

Useful6/10
Difficulty6/10
Novelty7/10
Paper: Entropy and domination for quasi-Hitchin representations arXiv:2608.27939
Unverified 2026

Concatenated-Neighbor Low-Rank Message Passing

Replace a sparse graph layer's separate edge transformations with one joint low-rank factorization of all transformations entering each target node. For target node i, concatenate the neighbor matrices horizontally, project all neighbor features into a shared low-dimensional receiving basis, and reconstruct one output; retain the self transformation exactly. This can reduce edge-parameter storage and message-passing FLOPs when the incoming block row has rapidly decaying singular values.

Useful6/10
Difficulty5/10
Novelty6/10
Paper: Accelerated S-NFC for Million-Chaff RCS Computation Using Low-Rank Compression of Concatenated Block Rows arXiv:2608.27936
Unverified 2026

Water-Filled Block-Sparse Neural Connectivity

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…

Useful6/10
Difficulty5/10
Novelty7/10
Paper: Graphon Design for Human-Machine Coordination under Bounded Rationality: Optimality of Stochastic Block Models arXiv:2608.27851
Unverified 2026

Uniform spectral graph coreset

Replace random node or token subsampling by a subset selected to preserve the averages of low-frequency graph signals exactly or approximately. Build a graph over nodes or tokens, compute its Laplacian eigenspace, and choose a fixed-cardinality subset whose eigenvector moments match those of the full graph; use the retained subset for graph pooling or downstream attention.

Useful6/10
Difficulty5/10
Novelty7/10
Paper: Uniformly Weighted Graphical Designs arXiv:2608.27624
Unverified 2026

Eager Masked Multiresolution Discovery

Replace recursive adaptive coordinate traversal in a CPPN-based architecture generator with one static multiresolution candidate tensor and an active boolean mask. Evaluate all candidate coordinates in parallel, compute local CPPN-output variance over each parent cell's children, and retain only children whose parent is active and whose predicted variation exceeds a threshold. This preserves adaptive refinement while enabling population batching, JIT compilation, GPU vectorization, and…

Useful6/10
Difficulty4/10
Novelty5/10
Paper: Tensor-Accelerated Eager Multi-Resolution Grids for Evolving Large-Scale Substrates arXiv:2608.27612
Unverified 2026

Conservative multifractal cascade prior

Use a conservative multiplicative cascade as a hierarchical latent prior or data-augmentation mechanism for models that generate intermittent, heavy-tailed, multiscale fields. The model receives a controllable cascade-width parameter, allowing systematic conditioning and evaluation across levels of non-Gaussianity instead of relying only on Gaussian latent noise.

Useful6/10
Difficulty5/10
Novelty6/10
Paper: Exact topology of conservative multiplicative cascades: An ultrametric transfer-operator genus arXiv:2608.24897
Unverified 2026

Sparse Levy Skip Network

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…

Useful6/10
Difficulty5/10
Novelty7/10
Paper: Universality of superdiffusion in simple random graphs arXiv:2608.23207
Unverified 2026

Cost-Map Finite-Action Head

Replace online enumeration over a finite action set with a classifier or lookup map whose regions directly return the action minimizing a one-step predictive-control cost. For affine dynamics and quadratic tracking loss, exact action regions are separated by pairwise cost boundaries, so the approximation can be audited against exhaustive predictive control rather than treated as an unconstrained policy.

Useful6/10
Difficulty5/10
Novelty6/10
Paper: A Simple and Extremely Efficient Predictive Control for Power Converters arXiv:2608.22416