Unverified
2026
Replace independent top-k routing by a tree-structured hypergraph assignment layer. Each candidate route is a singleton or pair of resources, and the router selects exactly q_e routes for every tree edge e while ensuring that no resource is consumed twice. This removes capacity collisions before expert computation instead of repairing them with token dropping or load-balancing penalties.
Useful6/10
Difficulty7/10
Novelty7/10
Unverified
2026
Replace a stationary federated optimizer with a decentralized optimizer whose target distribution explicitly forgets old streaming samples. Each round performs only K consensus-gradient iterations, with K selected from the mixing contraction so that the communication budget matches the temporal volatility of the objective. The method should react faster to distribution shifts while limiting disagreement and bias caused by heterogeneous clients.
Useful6/10
Difficulty5/10
Novelty5/10
Unverified
2026
Replace count-only MoE load balancing with greedy balancing of aggregate token-feature vectors. A token is assigned to the expert for which adding its feature vector produces the smallest increase in that expert's squared aggregate norm, encouraging experts to receive complementary semantic mixtures rather than identical token counts.
Useful6/10
Difficulty5/10
Novelty6/10
Unverified
2026
Grow a mixture-of-experts layer by splitting one expert into two children while conserving its routing mass, and choose the split ratio to minimize the worst imbalance over all intermediate expert counts. Use the paper's sharp threshold as a hard design target: with n experts, some intermediate stage must have capacity ratio at least D_n = 2^{1-1/\lceil n/2\rceil}, so schedules substantially below this are impossible rather than merely difficult to discover. Initialize child router logits with…
Useful6/10
Difficulty5/10
Novelty8/10
Unverified
2026
Replace a continuous allocation or routing decision with a lattice-valued decision whose unit size is explicitly normalized by total capacity. Round allocations downward rather than to the nearest lattice point, preserving per-example capacity feasibility, and train or evaluate against the resulting granularity ratio rather than treating discretization as an implementation detail.
Useful6/10
Difficulty5/10
Novelty7/10
Unverified
2026
Replace backpropagated gradients in a Langevin sampler with a simultaneous-perturbation estimate obtained from two noisy evaluations of a neural energy or simulator-defined negative log-density. This enables posterior or latent-space sampling when the energy contains nondifferentiable code, stochastic simulation, discrete operations, or an inaccessible neural-network gradient.
Useful6/10
Difficulty5/10
Novelty8/10
Unverified
2026
Add a small constrained equilibrium layer whose response depends on the current neural state and recent exogenous history, then cache responses keyed by a learned history embedding. For a new history, reuse a cached response only when an empirical Wasserstein distance to the cached history is below a threshold; otherwise run a few inner optimization iterations. The paper's local Holder and trajectory-stability results motivate graceful degradation rather than catastrophic errors for nearby…
Useful6/10
Difficulty5/10
Novelty6/10
Unverified
2026
Replace a fixed maximum-wait timeout in dynamic neural inference batching with a service-aware timer. The timer estimates the marginal reduction in computation or communication cost from adding another request and delays service only while that saving justifies the additional oldest-request delay.
Useful6/10
Difficulty4/10
Novelty6/10
Unverified
2026
Replace fixed-period federated averaging or distributed all-reduce with a Bernoulli communication decision whose probability is selected from estimated network connectivity and optimization conditioning. Local workers continue making corrected updates between communication events, while the contraction theorem exposes when communication is worth its cost.
Useful6/10
Difficulty5/10
Novelty6/10
Unverified
2026
Represent stochastic layer execution, branching, retries, and early exit as a finite continuous-time Markov chain, with the completed-prediction state absorbing. Learn transition rates jointly with neural-network weights, but use MFPT sensitivities to allocate rate changes according to their available control budget rather than allowing one routing edge to dominate halting-time control.
Useful6/10
Difficulty6/10
Novelty7/10
Unverified
2026
Replace uniform or purely loss-driven update allocation with a scheduler that targets both the mean update rate and the temporal variance of updates for each parameter group, task, or expert. At every training step, assign the available minibatch slots or accelerator workers to groups with the largest weighted deficits, preventing starvation while avoiding highly bursty update streams that can produce optimizer oscillations.
Useful6/10
Difficulty5/10
Novelty7/10
Unverified
2026
Add a periodic coarse optimization phase to SGD or Adam that operates on a compressed parameterization and returns a prolongated correction to the full network. Retain nonsmooth constraints or regularizers explicitly through a primal-dual update instead of relying on penalty smoothing. Accept the correction only when it improves a cheap fine-batch merit test, making the method useful even when the coarse objective is only approximately coherent with the fine objective.
Useful6/10
Difficulty6/10
Novelty7/10
Unverified
2026
Replace Euclidean projection onto a smooth convex parameter constraint by a local tangent-halfspace linear minimization step. The method is particularly suitable for convex softmax heads or low-dimensional adapter parameters constrained to an ellipsoid, where projections require matrix solves or expensive nonlinear routines but membership and boundary-normal queries are cheap.
Useful6/10
Difficulty5/10
Novelty7/10
Unverified
2026
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.
Useful6/10
Difficulty5/10
Novelty8/10
Unverified
2026
Replace communicated floating-point gradients in synchronous federated or data-parallel training with one-bit threshold queries whose thresholds are sampled publicly before gradients are observed. Use several fixed geometric amplitude scales so the same protocol handles unknown gradient means and heavy-tailed client updates without an interactive localization round. Decode each coordinate from the scale whose neighboring estimates are statistically consistent, then apply the decoded aggregate…
Useful6/10
Difficulty6/10
Novelty6/10
Unverified
2026
Replace heuristic magnitude pruning in a layerwise convex reconstruction problem with safe screening based on a perspective relaxation of the cardinality constraint. A channel is removed only when a lower bound for every solution containing that channel exceeds the loss of a feasible incumbent; conversely, a channel is forced to remain when every solution excluding it is provably worse.
Useful6/10
Difficulty5/10
Novelty6/10
Unverified
2026
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.
Useful6/10
Difficulty5/10
Novelty7/10
Unverified
2026
Use the decision diagram as an exact optimizer over feasible binary gate paths to separate strong inequalities at the current fractional architecture. Add only violated cuts to the LP or MILP relaxation, rather than enumerating all gate configurations or relying on weak pairwise product constraints.
Useful6/10
Difficulty7/10
Novelty8/10
Unverified
2026
Replace independent linearizations of binary neural-network gates and their higher-order interactions with a compact decision-diagram flow formulation. This preserves the exact convex hull of feasible gate configurations whenever the DD is exact, making MILP-based pruning, quantization, or architecture search substantially less vulnerable to fractional gate solutions.
Useful6/10
Difficulty6/10
Novelty7/10
Unverified
2026
Build a sparse expert cascade whose router uses empirical progress, uncertainty, and compute cost to construct an efficient frontier of experts. Instead of always choosing the expert with the largest immediate gain per FLOP, route different confidence states through an ordered sequence of frontier experts, allowing cheap high-variance experts early and safer or more cost-effective experts near the final decision.
Useful6/10
Difficulty5/10
Novelty6/10
Unverified
2026
Replace a costly full-data conjugate-gradient solve for a neural-network linearized least-squares step with a sequence of progressively larger sketched solves. Each solve starts from the previous solution, so early iterations cheaply identify the useful update direction and only the final few iterations use the full training batch.
Useful6/10
Difficulty5/10
Novelty6/10
Unverified
2026
Replace node-restricted graph prototypes by centers that may lie at arbitrary coordinates on graph edges. For each prototype, hold its assigned points and shortest-path branches fixed, then update its edge coordinate by an exact clipped weighted least-squares formula; recompute branch changes only at arc bottleneck points. This creates a graph-aware vector-quantization or retrieval module that can represent semantic locations between existing nodes.
Useful6/10
Difficulty5/10
Novelty6/10
Unverified
2026
Replace ordinary sign-gradient or fixed-range gradient quantization in distributed training with a non-adaptive one-bit refinement block. Workers generate all localization and multiscale residue bits using public randomness before the server knows the current aggregate gradient; after decoding a coarse center, the server chooses which stored phase and scale interpretation to use. This should reduce clipping and variance for heavy-tailed gradients without requiring a second communication round.
Useful6/10
Difficulty6/10
Novelty7/10
Unverified
2026
Train a primal neural PDE solver and a separate physical-adjoint neural solver, then use their first-order-system residuals to adaptively allocate collocation points toward regions that control a chosen quantity of interest. Instead of minimizing only the primal residual uniformly, prioritize points according to a balanced combination of primal and adjoint local residuals, because the target-output error is controlled by their global product.
Useful6/10
Difficulty5/10
Novelty6/10