papers checked against Recognition

60 of 33987 papers

• Frieze patterns in representation theory

  math.RT · 2025-09-20 · 3 theorem links

  A survey of results linking frieze patterns to polygon triangulations, Grassmannian cluster algebras, and Grassmannian cluster categories, with focus on recent links to cluster categories.

• A Derandomization Framework for Structure Discovery: Applications in Neural Networks and Beyond

  stat.ML · 2025-10-22 · 3 theorem links

  A derandomization lemma proves that optimizing E_x[g_θ(Wx + b)] converges to W = 0, explaining low-rank structure discovery in neural networks of arbitrary depth under weak regularization and any SOSP optimizer.

• Additive Rigidity for $x$-Coordinates of Rational Points on Elliptic Curves

  math.NT · 2025-10-04 · 3 theorem links

  If a d-dimensional proper GAP in Q contains a positive proportion ρ of the x-coordinates of rational points on an elliptic curve E of rank r, then the number of such points is at most A(E,d,ρ)^r.

• Quantum separability criteria from bipartite systems to multipartite systems based on generalized Bloch representation

  quant-ph · 2025-10-28 · 3 theorem links

  A parameterized extended correlation tensor from generalized Bloch representation yields separability criteria that detect bipartite and multipartite entanglement with improved performance.

• Syntomification and crystalline local systems

  math.NT · 2025-10-19 · 3 theorem links

  Equivalence of reflexive sheaves on syntomic stack X^Syn with Z_p-lattices in crystalline local systems on generic fiber X_η, plus results on etale realization and filtered F-isocrystals for proper smooth X.

• Gauge-Equivariant Graph Networks via Self-Interference Cancellation

  cs.LG · 2025-11-20 · 3 theorem links

  GESC is a gauge-equivariant GNN that uses self-interference cancellation via phase connections and rank-1 projections to improve performance on heterophilous graphs.

• Neural Policy Composition from Free Energy Minimization

  math.OC · 2025-12-04 · 3 theorem links

  Policy composition emerges from variational free energy minimization through a convergent gradient flow with a soft-competitive recurrent neural implementation.

• Boundary Calculus, Rigidity Islands, and Deformation Theory in Algebraic Phase Structures

  math.RA · 2026-01-26 · 3 theorem links

  Introduces boundary calculus and rigidity islands to stratify deformation behavior by boundary depth and failure type in algebraic phase structures.

• AgentClinic: a multimodal agent benchmark to evaluate AI in simulated clinical environments

  cs.HC · 2024-05-13 · 3 theorem links

  AgentClinic is a multimodal agent benchmark demonstrating that LLM diagnostic accuracy on MedQA drops to below one-tenth in sequential clinical simulations, with Claude-3.5 leading and large tool-use differences across models.

• Beyond Forgetting: Machine Unlearning Elicits Controllable Side Behaviors and Capabilities

  cs.LG · 2026-01-29 · 3 theorem links

  Representation misdirection unlearning guided by the linear representation hypothesis elicits controllable emergent behaviors and enhanced capabilities in LLMs beyond simple forgetting.

• From complex to non-Archimedean geometry: an approach to the YTD conjecture

  math.AG · 2026-02-11 · 3 theorem links

  Sketches a proof of a version of the Yau-Tian-Donaldson conjecture for constant scalar curvature Kähler metrics by relating algebraic, analytic, and non-Archimedean geometries.

• Generic Long-Range Order-Parameter Correlations in Metallic Quantum Magnets

  cond-mat.str-el · 2026-02-26 · 3 theorem links

  Coupling of order parameter to conduction electrons produces long-ranged susceptibility at T=0 in metallic magnets, generically driving first-order quantum phase transitions in most 3D homogeneous-magnetization systems while leaving some classes second-order.

• Flowette: Flow Matching with Graphette Priors for Graph Generation

  cs.LG · 2026-02-27 · 3 theorem links

  Flowette combines flow matching with graphette structural priors, optimal transport coupling, and regularization to generate graphs with improved topology, achieving competitive or state-of-the-art results on synthetic and molecular benchmarks.

• Multidimensional cost geometry

  math.DG · 2026-04-08 · 3 theorem links

  Multidimensional reciprocal cost function yields rank-one degenerate Hessian geometry in log coordinates and nondegenerate pseudo-Riemannian metric in x-coordinates, with affine/Levi-Civita geodesic comparison and divergence realizations.

• Parallelized Hierarchical Connectome: A Spatiotemporal Recurrent Framework for Spiking State-Space Models

  q-bio.NC · 2026-04-01 · 3 theorem links

  PHC upgrades SSMs to spatiotemporal recurrent networks, instantiated as the compact PHCSSM spiking model that integrates five biological priors and matches larger baselines on long sequences with cross-hardware reproducibility.

• CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery

  cs.AI · 2026-04-02 · 3 theorem links

  CORAL is a new autonomous multi-agent evolution framework that achieves 3-10x higher improvement rates on 10 open-ended tasks with fewer evaluations than fixed baselines, including improving a kernel engineering score from 1363 to 1103 cycles.

• Sparse Contextual Coupling Reshapes Diffusion Geometry in Multilayer Hypergraphs

  physics.soc-ph · 2026-05-19 · 3 theorem links

  Sparse contextual layers in multilayer hypergraphs substantially reshape diffusion distances and community structure in gene networks even when involving less than 2% of genes.

• The Causal Bootstrap: Bounding Smeared Spectral Functions from Non-Perturbative Euclidean Data

  hep-lat · 2026-05-19 · 3 theorem links

  The causal bootstrap computes rigorous bounds on smeared spectral functions from non-perturbative Euclidean data by optimizing over the convex set of compatible positive spectral densities and reducing dual problems to semidefinite programs for certain kernels.

• Quantitative exponential mixing for the randomized Chirikov standard map

  math.PR · 2026-05-20 · 3 theorem links

  Establishes quantitative exponential mixing for the randomized Chirikov standard map on T^2 under large kicking strengths via a new criterion for incompressible random dynamical systems.

• G{\aa}rding Polynomials

  math.CO · 2026-04-30 · 3 theorem links

  Gårding polynomials form a new class strictly extending real stable polynomials, with equivalent characterizations via polarization and partial derivatives, and applications to matroid generating functions yielding new negative dependence results.

• NOFE - Neural Operator Function Embedding

  cs.LG · 2026-05-12 · 3 theorem links

  NOFE is a neural operator method for continuous dimensionality reduction using Graph Kernel Operators that outperforms PCA, t-SNE and UMAP on local structure preservation and sampling independence in datasets including ERA5 climate reanalysis.

• Breaking $\textit{Winner-Takes-All}$: Cooperative Policy Optimization Improves Diverse LLM Reasoning

  cs.AI · 2026-05-12 · 3 theorem links

  GCPO uses team-level credit assignment via determinant volume over reward-weighted semantic embeddings to promote non-redundant correct reasoning paths, improving both accuracy and diversity in LLM training.

• SLASH the Sink: Sharpening Structural Attention Inside LLMs

  cs.AI · 2026-05-11 · 3 theorem links

  SLASH is a plug-and-play attention redistribution technique that counters attention sinks to enhance LLMs' intrinsic graph topology reconstruction without any training or fine-tuning.

• Tokenizer Fertility and Zero-Shot Performance of Foundation Models on Ukrainian Legal Text: A Comparative Study

  cs.CL · 2026-05-14 · 3 theorem links

  Tokenizer fertility varies 1.6x across models on Ukrainian legal text, Qwen uses 60% more tokens than Llama-family models, zero-shot outperforms few-shot by up to 26 points, and pre-war classifiers lose 27.9 points on invasion-era decisions.

• Sharp Spectral Thresholds for Logit Fixed Points

  cs.LG · 2026-05-15 · 3 theorem links

  For finite-dimensional affine logit systems the sharp dimension-free stability threshold is β‖ΠWΠ‖_{T→T}<2, extending the certified regime beyond classical conservative bounds.

• Amplitude Noise Suppression in Frequency-Doubled Lasers: A Lyapunov Mechanism for Intensity Stabilization in Coupled Oscillator Systems

  physics.optics · 2026-05-13 · 3 theorem links

  A Lyapunov functional in the chi2 dynamics of intracavity frequency-doubled lasers establishes a constant-intensity manifold that suppresses amplitude noise, experimentally confirmed by 100-fold contrast in an Nd:YVO4-LBO system.

• Combinatorial Approach to the Second Law

  math.CO · 2026-05-14 · 3 theorem links

  A combinatorial study of mechanisms producing irreversibility from reversible dynamics, framed as an approach to the second law.

• MedMIX: Modality-Internal Expert Fusion for Multimodal Medical Diagnosis

  cs.LG · 2026-05-15 · 3 theorem links

  MedMIX combines intra-modality expert fusion, learned inter-modality fusion, and training-only large-small collaboration to deliver robust multimodal medical prediction under incomplete modalities across three benchmarks.

• Whitney's 2-isomorphism theorem for graphings

  math.CO · 2026-05-15 · 3 theorem links

  Measurable versions of Whitney's 2-isomorphism theorem are established for locally finite graphings by defining weak isomorphisms that preserve edge measures, cycles, and hyperfinite subgraphs, with rigidity for weakly 3-connected infinitely-ended cases and implementation via countable measurable Wh

• Five legs @ three loops: N=4 sYM amplitude near mass-shell

  hep-th · 2026-05-16 · 3 theorem links

  Three-loop five-leg amplitude in planar N=4 sYM near mass shell is computed via 6D unitarity cuts and dimensional reduction, confirming IR exponentiation governed by octagon anomalous dimension with each of three kinematic structures having its own function of 't Hooft coupling.

• Fidelity Probes for Specification--Code Alignment

  cs.LG · 2026-05-17 · 3 theorem links

  Fidelity probes from code raise specification fidelity from 0.63 to 0.94 on a 12k-line COBOL benchmark over eight iterations, with convergence predicted by a two-state Markov fixed point from four iterations of rate data.

• ChemVA: Advancing Large Language Models on Chemical Reaction Diagrams Understanding

  cs.AI · 2026-05-17 · 3 theorem links

  ChemVA framework uses hybrid-granularity visual anchors and entity-name alignment to improve LLM performance on chemical reaction diagrams by ~20 points, reaching 92% structural accuracy on the new OCRD-Bench dataset.

• Scale-Equivariant Generative Forecasting: Weight-Tied Dilated Convolutions, Wavelet Scattering Inputs, and Spectral-Consistency Training for Self-Similar Time Series

  cs.LG · 2026-05-17 · 3 theorem links

  Presents SE-WaveNet with weight-tied dilated convolutions plus wavelet and spectral components that reproduces empirical scaling collapse on financial returns while using L times fewer convolutional parameters.

• Do generative video models understand physical principles?

  cs.CV · 2025-01-14 · 3 theorem links

  Physics-IQ benchmark reveals that generative video models exhibit limited physical understanding unrelated to their visual quality.

• ScheduleFree+: Scaling Learning-Rate-Free & Schedule-Free Learning to Large Language Models

  cs.LG · 2026-05-18 · 3 theorem links

  ScheduleFree+ scales schedule-free learning to LLMs with fixes for large batches and models, outperforming Warmup-Stable-Decay schedules by up to 31% at 1000 tokens per parameter.

• How Far is Video Generation from World Model: A Physical Law Perspective

  cs.CV · 2024-11-04 · 3 theorem links

  Video generation models generalize perfectly inside the training distribution but fail out-of-distribution and rely on case-based mimicking of nearest training examples instead of abstracting physical laws.

• FAGER: Factually Grounded Evaluation and Refinement of Text-to-Image Models

  cs.CV · 2026-05-18 · 3 theorem links

  FAGER is a new agentic framework that creates structured factual rubrics to evaluate and refine text-to-image outputs for implicit factual correctness across science, history, products, and culture.

• Position: A Three-Layer Probabilistic Assume-Guarantee Architecture Is Structurally Required for Safe LLM Agent Deployment

  cs.AI · 2026-05-18 · 3 theorem links

  A three-layer probabilistic assume-guarantee architecture is structurally required for safe LLM agent deployment.

• Quantum--Fluid Correspondence for Systems of Nonrelativistic Spin-$\frac{1}{2}$ Particles

  quant-ph · 2026-05-18 · 3 theorem links

  Derives the Pauli equation from a charged spinning fluid model and reformulates n-particle systems as 3n-dimensional Euler flows.

• OrganicHAR: Towards Activity Discovery in Organic Settings for Privacy Preserving Sensors Using Efficient Video Analysis

  cs.HC · 2026-05-18 · 3 theorem links

  OrganicHAR discovers 4-8 activity categories per user from sensor signals, achieves 79% accuracy on coarse activities with ambient sensors alone and cuts VLM queries by 90% by triggering video analysis only at detected pattern moments.

• On nematic electrolytes

  math.AP · 2026-05-18 · 3 theorem links

  Proves existence, uniqueness, global existence criteria, and equilibrium characterization for strong solutions to a Nernst-Planck-Poisson-Navier-Stokes-Ericksen-Leslie system modeling nematic electrolytes via maximal regularity.

• TurboQuant: Online Vector Quantization with Near-optimal Distortion Rate

  cs.LG · 2025-04-28 · 3 theorem links

  TurboQuant achieves near-optimal vector quantization distortion for both MSE and inner products via random rotation and per-coordinate scalar quantization, with a formal proof that it matches lower bounds within a factor of approximately 2.7.

• DiffuSeq: Sequence to Sequence Text Generation with Diffusion Models

  cs.CL · 2022-10-17 · 3 theorem links

  DiffuSeq adapts diffusion models to conditional sequence-to-sequence text generation and reports performance matching or exceeding strong baselines including pretrained language model systems while generating more diverse outputs.

• Beyond Isotropy in JEPAs: Hamiltonian Geometry and Symplectic Prediction

  cs.LG · 2026-05-19 · 3 theorem links

  Fixed isotropic marginals in JEPAs can be maximally misaligned with unknown structured geometries, and HamJEPA using symplectic Hamiltonian leapfrog maps improves kNN and linear-probe performance on CIFAR-100 and ImageNet-100.

• CADENet: Condition-Adaptive Asynchronous Dual-Stream Enhancement Network for Adverse Weather Perception in Autonomous Driving

  cs.CV · 2026-05-19 · 3 theorem links

  CADENet introduces an asynchronous dual-stream enhancement network with CAPE and EG-NMS plus CLIP zero-shot classification to improve camera-based detection in rain, fog, snow and sand without retraining or latency penalty, reporting low F1 scores on DAWN as lower bounds due to annotation bias.

• Drifting Objectives for Refining Discrete Diffusion Language Models

  cs.CL · 2026-05-19 · 3 theorem links

  TokenDrift refines discrete diffusion language models by applying anti-symmetric drifting to soft-token features during training, yielding large reductions in generation perplexity at low NFEs.

• Unveiling Energetic Advantage in Superconducting Cat-Qubits Quantum Computation

  quant-ph · 2026-05-19 · 3 theorem links

  Energy modeling and parameter optimization for cat-qubit superconducting quantum computers performing semiclassical QFT with error correction indicates an energetic advantage over classical systems for more than 26 qubits under cryogenic assumptions.

• Stochastic trajectories and excursions in a double quantum dot system

  quant-ph · 2026-05-19 · 3 theorem links

  Uses stochastic excursions to compute averages, noise, and uncertainty relations for transport observables in a double quantum dot, highlighting trade-offs in successful versus unsuccessful events.

• Genus formulas for dormant modular curves and asymptotic behavior of their function fields

  math.AG · 2026-05-18 · 3 theorem links

  Derives explicit genus formulas for dormant modular curves from dormant PGL2-opers and analyzes asymptotic behavior of their function field towers.

• On efficient robust regression with subquadratic samples

  cs.DS · 2026-05-18 · 3 theorem links

  Near-linear time algorithm for robust regression under Gaussian covariates achieves O(sqrt(ε κ)) error with Õ(d/ε⁴) samples when ε κ ≲ 1, plus SQ and low-degree lower bounds.

• Distributed Renaming with Subquadratic Bits via Scalable Committee Election

  cs.DC · 2026-05-18 · 3 theorem links

  Two Byzantine-tolerant renaming algorithms using a new scalable committee election primitive achieve O(polylog n) time and Õ(n) or Õ(n + min{nf,T}) communication.

• Shapiro steps of superfluid Fermi gases in a ring trap across the BCS--BEC crossover

  cond-mat.quant-gas · 2026-05-18 · 3 theorem links

  Time-dependent BdG simulations show low-order Shapiro steps in barrier velocity versus chemical potential difference for ring-trapped superfluid Fermi gases, with quantization in units of ħω/2.

• Decision-Aware Proximal Bridge Learning for Optimal Treatment Selection

  cs.LG · 2026-05-16 · 3 theorem links

  Introduces decision-aware proximal bridge learning using a weighted loss and regret bound to enhance optimal treatment selection in settings with hidden confounding.

• Beyond Point-Wise Matching: Structural Representation Alignment for Accelerating Diffusion Transformers

  cs.CV · 2026-05-16 · 3 theorem links

  sREPA enforces structural consistency in relational geometry of pre-trained vision features to accelerate DiT training and improve generation quality.

• Shifted quantum toroidal algebra of type $\mathfrak{gl}_{1|1}$ and the Pieri rule of the super Macdonald polynomials

  math.QA · 2026-05-16 · 3 theorem links

  Super Macdonald polynomials indexed by super partitions form a basis of the level zero super Fock module of the shifted quantum toroidal algebra U_{q,t}(gl hat hat 1|1), with the Pieri rule following from super charge actions and yielding supersymmetric Hamiltonians via anti-commutators.

• Representing Higher-Order Networks: A Survey of Graph-Based Frameworks

  cs.SI · 2026-03-24 · 3 theorem links

  A comprehensive survey of graph-based frameworks for higher-order networks, covering foundational concepts, extensions, and newly introduced formalisms with emphasis on structural principles and applications.

• The Geometric Structure of Models Learning Sparse Data

  cs.LG · 2026-05-08 · 3 theorem links

  Normal alignment is the rank-one Jacobian structure that lets classifiers minimize loss and maximize local robustness in sparse regimes; the paper proves its optimality and uses it to create GrokAlign and RFAMs.

• Scalable neuromorphic computing from autonomous spiking dynamics in a clockless reconfigurable chip

  cs.NE · 2026-05-15 · 3 theorem links

  Clockless FPGA circuits produce autonomous spiking neuron networks that achieve competitive audio classification accuracy with significantly lower power than conventional digital implementations.

• $\phi$-Balancing for Mixture-of-Experts Training

  cs.LG · 2026-05-14 · 3 theorem links

  φ-balancing is a convex optimization method for population-level expert balance in MoE training that derives an online EMA adjustment and outperforms heuristic baselines.

• ScienceBoard: Evaluating Multimodal Autonomous Agents in Realistic Scientific Workflows

  cs.AI · 2025-05-26 · 3 theorem links

  ScienceBoard creates a new environment and 169-task benchmark for testing multimodal agents on realistic scientific workflows using real software, revealing a 15% success rate for state-of-the-art models.

next page →

← previous page