Peer Revenue Now accepting positions

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Friday 6 November

Analysis of PDEs

Dynamics of subcritical threshold solutions for the three-dimensional energy-critical NLS

We extend the radial threshold classification of Duyckaerts and Merle on the lower kinetic branch to nonradial solutions of the three-dimensional focusing energy-critical nonlinear Schrödinger equation in \(\dot H^1(\mathbb R^3)\). At the ground state energy and with initial gradient norm strictly below that of the ground state, every solution is global and either scatters in both time directions or is a symmetry transform of the special solution \(W^-\), possibly after conjugate time reversal. The proof combines Su and Zhao's below-threshold scattering theorem with a coercivity estimate for the interaction form at threshold under a kinetic gap and a local analysis near the ground state orbit.
Read the full abstract

Information Theory

Enhanced Informed Dynamic BP Decoding Scheduling Strategies for 5G NR LDPC Codes

We present an enhanced informed dynamic scheduling (IDS) strategy to avoid propagations of unreliable messages in the residual belief-propagation (RBP) algorithm for decoding low-density parity-check (LDPC) codes. To reduce the resource needed for updating the degree-one variable nodes (VNs) in an RBP-based algorithm for decoding 5G New Radio (NR) LDPC codes, we further introduce a process consisting of several simple procedures for managing the resource consumption. Simulation results demonstrate that our RBP algorithm based on the Proposed IDS strategy outperforms existing RBP-based algorithms at the early decoding iterations. By using our degree-one VN resource managing approach, the performance of both the proposed and the existing decoders is significantly improved in decoding the 5G NR codes.
Read the full abstract

All authors Tofar C.-Y. Chang, I-Hsiang Lee, Pin-Han Wang, Jian-Jia Weng, Yu T. Su

Computational Complexity

On derandomizing the line-point low-degree test

We investigate the derandomized version of the classical line-point low-degree test in the high-error regime in the bivariate setting and reveal some unexpected behavior. The classical line-point low-degree test takes an $m$-variate function $f$ over a finite field $\mathbb{F}_q$, queries it along a random line and accepts if the restriction of the function to this line is a low-degree polynomial. The high-error version of (the analysis of) this test asserts that if $f$ has at least $\epsilon$ expected local agreement with a degree-$k$ univariate polynomial on a random line, then in fact $f$ is non-trivially close to some degree-$k$ $m$-variate polynomial even as $\epsilon$ approaches $0$ (provided $q$ is sufficiently large compared to $k$ and $1/\epsilon$). The derandomized version of the problem asks if the same holds when the line is not chosen uniformly over all possible lines but rather from a small set. In this paper we explore the bivariate case ($m=2$). We show, somewhat surprisingly, that there exists a function $f\colon \mathbb{F}_q^2 \to \mathbb{F}_q$ and a set of lines $\mathcal{L}$ in $\mathbb{F}_q^2$ of size nearly $q^{3/2}$ such that on every line in $\mathcal{L}$, $f$ has agreement $\approx 1/2$ on average with some univariate polynomial of degree $k=\sqrt{q}/2$ and yet $f$ is $1 - O(q^{-1/2})$ far from every polynomial of degree $O(k)$. On the positive side, we show that once the field size $q$ is at least $k^2/(\epsilon)^{O(1)}$ such examples disappear and we can get sets of lines of size $O(q/\epsilon^{O(1)})$ on which the low-degree test has a high-error analysis. On the way to our positive result, we show a new high-degree version of the classical two-direction Polishchuk–Spielman divisibility lemma.
Read the full abstract

All authors Prahladh Harsha, Swastik Kopparty, Mrinal Kumar, Ramprasad Saptharishi, Madhu Sudan

Quantum Physics

A Tutorial on Practical QKD Post-Processing Algorithms Using a GUI-Based Simulation Framework

Recent experimental demonstrations of Quantum Key Distribution (QKD) have attracted significant attention for their ability to establish information-theoretically secure encryption keys for next-generation communication systems. As practical deployment accelerates, the realization of compact, scalable, and high-throughput discrete-variable (DV) QKD systems has become increasingly important. While advances in photonic integration are enabling low-size, low-weight, and low-power QKD hardware, the security and performance of these systems rely heavily on efficient post-processing techniques to generate identical, secure keys. This tutorial presents a comprehensive algorithm-level review of practical post-processing techniques for DV-QKD, with particular emphasis on information reconciliation methods based on Cascade, Winnow, low-density parity-check (LDPC), and polar codes. We also analyze the theoretical asymptotic and finite-size secret key rates (SKRs) of DV-QKD and benchmark these bounds against the practical SKRs achieved using actual post-processing algorithms. Furthermore, we present a server-client model to implement the entire QKD post-processing pipeline. The performance is evaluated in terms of QBER and SKR across various system parameters, considering the complete post-processing pipeline, including sifting, information reconciliation, privacy amplification, and authentication. This tutorial also presents a graphical user interface (GUI) for the QKD post-processing methods, enabling researchers, industry practitioners, and academic users to visualize and analyze the impact of different post-processing techniques and system parameters, such as QBER, on the SKR performance of DV-QKD systems.
Read the full abstract

All authors Ushnik Nath, Neha Sharma, Ramniwas Meena, Neel Kanth Kundu

Dynamical Systems

Hecke Algebras for Minimal Equicontinuous Cantor Actions

In this work we extend the concept of Hecke algebras to a pair of a separable infinite profinite group and a closed core-free subgroup. Motivated by applications to minimal equicontinuous Cantor actions, which are characterized by such pairs, we characterize isometric *-algebra isomorphisms between such Hecke algebras by showing that they always arise as Koopman operators of homeomorphisms between the double coset spaces combined with the multiplication by some character. We then prove that two minimal equicontinuous Cantor actions with isometric isomorphic Hecke algebras have the same Steinitz order. We provide examples including the infinite dihedral group and the Grigorchuk group.
Read the full abstract

Pattern Formation and Solitons

Continuum mean-field theory of spatiotemporal pattern formation in a three-state active Potts model

We investigate the spatiotemporal dynamics of a three-state active Potts model with a cyclic transition bias. Starting from microscopic Glauber-type dynamics, we derive macroscopic continuum field equations through a spatial mean-field approximation. Linear stability and weakly nonlinear analyses reveal that increasing the coupling strength induces a Hopf bifurcation, near which the critical amplitude dynamics is described by the complex Ginzburg–Landau equation. In the spatially homogeneous limit, the system exhibits limit-cycle oscillations and ordered states, with an intermediate regime of bistability bounded by a $\mathbb{Z}_3$-symmetric heteroclinic bifurcation. In two dimensions, the system sustains stable rotating spiral waves. Using perturbation theory based on one-dimensional traveling fronts, we semianalytically predict the curvature-dependent propagation velocity of circular fronts, as well as the rotational angular velocity and pitch of spiral waves. Finally, we compare the continuum model with discrete Monte Carlo simulations to clarify the role of thermal fluctuations and evaluate the validity of the mean-field approximation.
Read the full abstract

All authors Yuta Tateyama, Songhao Yin, Hiroaki Ito, Hiroyuki Kitahata, Hiroshi Noguchi

Analysis of PDEs

Positive self-similar solutions of a parabolic equation with spatially weighted absorption in a half-space

We study positive solutions of $u_t-\Delta u+x_1^{2\alpha}|u|^{q-1}u=0$ in the half-space $\{x_1>0\}$, with $\alpha>-1$ and $q>1$, vanishing on $\{x_1=0\}$ and at $t=0$ away from the origin. Let $q_c=(N+3+2\alpha)/(N+1)$. For $1<q<q_c$ we prove that there is exactly one positive self-similar solution continuous away from the origin, and that its profile satisfies Gaussian bounds that are sharp up to multiplicative constants. For $q\ge q_c$ there is none. We also show that, after removing the Gaussian factor, the profile has an angular limit shape $\Phi$ in every mode, give an integral formula for its Kummer coefficients, prove that $\Phi$ is real-analytic and is the uniform limit of the rescaled profiles, that $\Phi/(\eta_1/|\eta|)$ is monotone along meridians, and that $\Phi$ is proportional to $\eta_1/|\eta|$ if and only if $2\alpha+q=1$. Every self-similar solution of finite energy, or continuous away from the origin, is dominated in absolute value by the positive one, and for $(N+4+2\alpha)/(N+2)\le q<q_c$ the only nontrivial ones are $\pm u_s$. For $\alpha=0$ existence, uniqueness and the upper bound are due to Marcus and Véron; the lower bound with the sharp exponent and the description of $\Phi$ are new also in that case. The existence and uniqueness proofs use a Brezis–Oswald type comparison principle and explicit barriers, given by explicit formulas; the analysis of $\Phi$ uses Kummer functions.
Read the full abstract

Machine Learning

A Taxonomy-Driven Synthetic Dataset for Retrieval Effectiveness and Fairness Evaluation in Hiring

While machine learning hiring tools are ubiquitous, the need for privacy regulations and commercial sensitivity makes public benchmarking legally complex. We address this challenge by introducing a taxonomy-driven, fully synthetic HR dataset and benchmark to simultaneously evaluate retrieval effectiveness and algorithmic fairness. Grounded in 660K real freelancer profiles, our multilingual corpus features candidate profiles, project briefs, and counterfactual demographic twins. Auditing state-of-the-art embedding models reveals performance divergence from general MTEB scores, alongside fairness disparities in grammatically gendered languages. This demonstrates that general benchmarks are insufficient for HR, requiring tailored evaluations to identify Pareto-optimal models that balance semantic accuracy and fairness.
Read the full abstract

All authors Warren Jouanneau, Morgane Hoffmann, Emma Jouffroy, Marc Palyart

Artificial Intelligence

Analyzing Code Provenance in LLMs for Programming Education

Large language models (LLMs) play an increasing role in programming education, challenging traditional assessment methods as they can generate fully functional code instantly. This disrupts the reliability of student submissions as indicators of learning and effort. To address this issue, code provenance is essential for tracing the origins of code and distinguishing student-authored, AI-generated, and externally sourced components. We extend the Min-K% probability approach by incorporating synthetic non-member code generation and apply it to Python code for membership inference and provenance tracking. Evaluations cover free LLMs such as GPT-2, DistilGPT-2, and DialoGPT, as well as state-of-the-art commercial models including GPT-5, Gemini 2.5, Mistral, DeepSeek, and LLaMA 3.3. Before evaluating our approach on higher-education student assignments, we validate its feasibility using Stack Overflow data. The results show that free models enable more sensitive code memorization detection than commercial models. By systematically comparing these LLMs, we assess trade-offs between detection accuracy and privacy protection, taking an initial step toward effective assessment of AI-assisted programming education through combined provenance detection and process visibility.
Read the full abstract

All authors Subhankar Maity, Faiza Belbachir, Omar Boucelma, Guilherme Medeiros Machado

Artificial Intelligence

TailDART: Compute-Efficient Foundation Models for Financial Time-Series Forecasting

Although time-series foundation models (TSFMs) achieve strong zero-shot transferability across high-information-density physical systems, applying overparameterized TSFMs to quantitative finance is fundamentally constrained by low signal-to-noise ratios, non-stationarity, and prohibitive autoregressive decoding latency. To reconcile representation learning with institutional operational constraints, this study presents TailDART, a dense 2.68 million-parameter foundation model engineered for financial time-series forecasting. By integrating grouped-query attention into a lightweight backbone, the architecture drastically reduces Key-Value cache memory bandwidth overhead, achieving major inference acceleration (up to 48.1%) over prevailing state-of-the-art TSFMs to enable low-latency, high-throughput streaming execution. Furthermore, multi-asset benchmarks across varying temporal resolutions show that a compact model can remain competitive for market forecasting: TailDART matches the predictive fidelity of substantially larger baselines while demonstrating superior error resilience in high-volatility regimes.
Read the full abstract

All authors Chiang Qin Zhi, Ranjan Satapathy, Erik Cambria, Joyjit Chattoraj

Machine Learning

SIFR: Stable Signed Flow Reinforcement

Reward-based fine-tuning of flow models requires effectively promoting high-reward and suppressing low-reward samples, while maintaining training stability over many rounds of iterative updates. Iterative reward fine-tuning provides a simple starting point, but the positive reward weights give no explicit unfitting signal, and repeated reweighting concentrates the distribution and causes collapse. We propose Stable Signed Flow Reinforcement (SIFR), which replaces the reward weights by signed advantages to supply the missing unfitting signal, and constrains the velocity drift from a reference model to stabilize the updates, with a Lagrange multiplier that adapts the regularization strength to a prescribed drift budget. We further show that existing flow RL methods fall within this framework: they implicitly minimize the signed loss, differing in their regularization choices. We validate SIFR on SD3.5 Medium (2.5B) and FLUX.2 Klein (9B) and scale it to Cosmos 3 Super Text2Image (64B): across all 12 backbone–reward pairs with preference and rule-based rewards, it matches or exceeds existing methods, reaches their peak rewards much faster, and keeps improving over long training where they plateau or collapse.
Read the full abstract

All authors Runlong Liao, Xingqian Xu, Lizhang Chen, Baiyu Su, Yufan Huang, Zhonghao Wang, Akash Gokul, Ming-Yu Liu, Qiang Liu, Yogesh Balaji

Numerical Analysis

Construction of DAE solver Rodas4PW and numerical test examples with Rosenbrock and W-methods in OrdinaryDiffEqRosenbrock.jl

This paper compares various Rosenbrock and Rosenbrock W-methods. To achieve this, some methods implemented in the Julia package OrdinaryDiffEqRosenbrock.jl are selected and benchmarked using test examples. The test examples include stiff ODE problems as well as differential-algebraic equations. In addition to the already known methods, the new W-method Rodas4PW is introduced and included in the tests.
Read the full abstract

Machine Learning

TOFS: Feature Selection with Swap Supervision for Semantic Communication over Fading Channels

Task-oriented semantic communication transmits task-relevant features to support receiver inference under limited communication resources and adverse channel conditions. For image classification, the usefulness of a feature depends on the other transmitted features and the distortion introduced by the wireless channel. Effective feature prioritization should therefore account for classification performance after transmission of the selected features. This paper proposes a Task-Oriented Feature Selection (TOFS) framework and method, which evaluates feature replacements using receiver classification loss. Loss differences confirmed under independent transmission samples provide pairwise ranking supervision for a selector network. The trained selector predicts priorities in a single forward pass. CIFAR-10 simulations demonstrate improved classification accuracy over an importance-based feature selection baseline at the same transmitted feature count and average transmit power. The proposed TOFS framework shows 84% accuracy performance under Rayleigh fading and Gaussian noise, compared with the 68% performance of a SOTA method.
Read the full abstract

All authors Zhiyi Zhu, Eiji Takimoto, Patrick Finnerty, Chikara Ohta

Data Structures and Algorithms

PaB-tree: Parallel B+-tree with Batch Updates

The B+-tree is a ubiquitous search index due to its provable I/O (cache) efficiency. Yet unlike batch-parallel binary search trees, which are both fast and widely available, no batch-parallel B+-tree offers strong empirical performance together with provable, near-optimal I/O efficiency. Prior B+-tree-based designs are either parallel but purely theoretical, with no high-performance implementation, or concurrent indexes such as the Bw-tree and Masstree, which run an order of magnitude slower on batch updates. We present PaB-tree, the first batch-parallel B+-tree with guarantees on work, span, and cache complexity alongside strong practical performance. Its core challenge – repairing node overflows and underflows without eroding the tree's I/O advantage – is resolved by one simple invariant: a repair never touches a sibling and is always pushed to the parent. A Promote primitive reduces batch construction and insertion to repeated node promotion, and batch deletion to reinsertion. Batch updates take O(B log_B n) work and O(log_B n) I/O per element, matching the sequential B+-tree bounds, with polylogarithmic span. Against highly-optimized parallel and concurrent baselines, PaB-tree delivers the best overall performance trade-off: up to 1.4x faster on construction and up to 3.6x faster on batch insertion than the strongest baseline, while preserving B+-tree query efficiency.
Read the full abstract

All authors Ziyang Men, Yan Gu, Yihan Sun, Helen Xu

Numerical Analysis

A structure-preserving discontinuous Galerkin method for variable-density incompressible flow with a partitioned time integrator

Preserving density bounds and mass conservation, together with a conservative momentum discretization and exact incompressibility, is a central challenge in high-order simulations of variable-density incompressible flow. We develop a structure-preserving discontinuous Galerkin method that couples conservative density transport with an \(H(\mathrm{div})\)-conforming mixed discretization of the momentum equation. The method uses a partitioned time integrator: density is advanced explicitly by strong-stability-preserving Runge–Kutta stages with a bound-preserving limiter, while momentum is updated by a linearly implicit backward differentiation formula. The transport velocities are extrapolated from previously computed mixed velocities. Because the discrete solenoidal space is linear, this extrapolation preserves elementwise zero divergence and single-valued normal traces, providing compatible transport velocities without an auxiliary reconstruction or projection. For source-free transport on triangular meshes with impermeable or periodic boundaries, we prove bounds for density cell averages and limiter control-point values under the stated spatial, quadrature, and stagewise Courant–Friedrichs–Lewy conditions. The conservative fluxes and mean-preserving limiter yield global mass conservation. Numerical experiments on triangular and structured rectangular meshes demonstrate the expected convergence rates, machine-level divergence errors, and bound preservation, including in a low-density regime with high density ratio.
Read the full abstract

Computer Vision

FIND: Frequency Invariance Disentanglement for Test-Time Adaptation in LiDAR 3D Detection

LiDAR-based 3D Object Detection (L3OD) is a fundamental 3D perception task, yet its performance is often compromised by domain shifts arising from diverse sensor configurations and environmental conditions. Existing Test-Time Adaptation (TTA) methods primarily leverage self-supervision in the spatial domain, yet they frequently confound semantic content with domain style due to the sparse, non-Euclidean nature of point clouds. To overcome this limitation, we propose FIND (Frequency INvariance Disentanglement), a novel TTA framework that precisely disentangles domain-invariant content from domain-specific interference within the frequency domain. Our methodology utilizes B-Spline fitting to construct locally adaptive filters integrated into a dual-stream architecture: an Invariant Stream extracts robust features to guide pseudo-labeling, while a Specific Stream enforces consistency against domain perturbations. By combining invariance-driven alignment with stability-guided regularization, our approach dynamically extracts robust domain-invariant features while suppressing domain-specific interference. Extensive experiments on cross-dataset adaptation and robust corruption benchmarks demonstrate that FIND significantly outperforms state-of-the-art methods. The source code is publicly available at https://github.com/llldx03/FIND.
Read the full abstract

Optimization and Control

The multiplier step size in classical two-block ADMM: convergence and counterexamples

How large can the multiplier step be in the classical two-block alternating direction method of multipliers (ADMM)? We keep the penalty fixed and use the usual primal updates. We consider finite-dimensional problems with closed proper convex objectives. We assume that a Karush-Kuhn-Tucker point exists and that each block subproblem attains its minimum at every step. Under these assumptions, the constraint images and the multiplier converge for a range of factors that extends beyond the golden ratio. The constraint matrices may be rank deficient. For larger steps, we construct two fixed compact polytopes and give an explicit threshold below two. The polytopes admit an infeasible periodic orbit for every factor at or above this threshold. At that threshold, ADMM can cycle even for smooth strongly convex objectives. This gives a negative answer to Glowinski's question about the full interval from zero to two. A different pair of piecewise-quadratic objectives gives geometrically growing residuals. For the general convex class, we locate the endpoint of the uniform convergence interval near 1.88688256. A convergence certificate on the unit circle gives the lower bound, while a separate periodic orbit gives the upper bound. Exact arithmetic verifies that the two bounds differ by less than 0.000000174.
Read the full abstract

Strongly Correlated Electrons

Lefschetz-Thimble Method to the Sign Problem in the Sachdev-Ye-Kitaev Model

The sign problem in path integrals severely limits the application of quantum Monte Carlo methods to strongly correlated systems, whereas the Lefschetz-thimble method offers a rigorous framework to mitigate this issue. In this paper, we establish a numerical framework based on the Lefschetz thimbles for the Sachdev-Ye-Kitaev (SYK) model. By introducing continuous auxiliary fields via the Hubbard-Stratonovich transformation, the partition function is formulated as a high-dimensional complex integral. We employ the complex Langevin equation to explore the saddle-point distribution of the effective action in the complex plane. Monte Carlo integration is then performed to compute the partition function and Green functions. Our numerical results exhibit excellent agreement with high-temperature expansions, infrared-limit solutions, and determinantal quantum Monte Carlo (DQMC) simulations, confirming the validity of our framework. Furthermore, we demonstrate that our complex Langevin search successfully captures the relevant dominant saddle points governing the SYK model in the large-N limit. Our work deepens the understanding of the sign problem in the SYK model and provides a viable non-perturbative avenue for addressing sign problems in more complex strongly correlated systems.
Read the full abstract

Neural and Evolutionary Computing

OMCDE: Opposition-based Multi-Stage Matrix-Based Competitive Differential Evolution

Differential Evolution has proven to be an effective algorithm for continuous global optimization, but it's modern high-performing variants often introduce additional complexity and computational cost. We propose OMCDE, a matrix-based competitive Differential Evolution algorithm that aims to improve optimization performance while maintaining computational efficiency. The proposed approach integrates competition-driven search dynamics, adaptive behavior between optimization stages, and opposition-based initialization within a parallelizable framework. Empirical results on widely used CEC 2017 functions show that OMCDE achieves strong performance compared to several state-of-the-art DE algorithms, while remaining efficient and scalable. These findings indicate that OMCDE is a promising alternative for solving complex numerical optimization problems.
Read the full abstract

Computational Geometry

Separating Measures by Hyperplanes

Given red and blue points in $\mathbb{R}^d$, connect every red point to every blue point and weigh each segment by its length under a fixed norm. What fraction of the total weight can always be captured by a single affine hyperplane separating the segments' endpoints? We study this question for finite point configurations in affine general position and, more generally, for pairs of probability measures. For dimension one, we show that choosing the best of three thresholds determined by the means of the two distributions captures at least one third of the total red-blue distance. We also give a matching upper bound. Embedding the line construction to higher dimensions shows that the guarantee of $1/3$ cannot be improved for arbitrary measures. Thus, we pivot to the family of measures whose atoms are in affine general position and whose non-atomic part places zero mass on all hyperplanes. Crucially, we establish that the worst-case guarantee is identical for finite configurations in affine general position and several natural classes of probability measures, thereby unifying the discrete and continuous formulations of the problem. For this family of measures and for higher dimensions, we extend our guarantee of $1/3$ for all norms. Our proofs combine the ham-sandwich theorem with a novel coupling argument and, in the planar case, an integral representation of norms. We also show that the optimal guarantee improves as the dimension grows, giving a lower bound of $1/2-\operatorname{O}\left(1/\sqrt{d}\right)$ as well as an asymptotically matching upper bound of $1/2$ for every $d$. For the Euclidean distance, we refine the lower bounds to $8/(8+5\pi)$ and $1/2-\operatorname{O}\left(1/d\right)$ and, for $d=2$, the upper bound to $5/11$. We also derive dimension-dependent guarantees for finite point configurations, including perfect separation of $n$ points of each color when $d\geq2n-1$.
Read the full abstract

All authors José Correa, Svenja M. Griesbach, Elfarouk Harb, Felix Hommelsheim, Reda Jlibene, Vasilis Livanos, Eilon Solan

Optics

Low-Rank Iteration for Efficient Calculations of Topological Surface States

Surface states are hallmarks of topological phases, and their efficient computation is crucial for topology characterization. Conventional supercell methods (SCM) approximate semi-infinite systems as finite periodic ones, but suffer from high computational cost, artificial inter-surface coupling, and slow convergence near band edges. Recently, boundary problems for semi-infinite systems have been recast as block-tridiagonal equations and solved by the cyclic reduction method (CRM) and the transfer matrix method (TMM). However, these approaches remain resource-intensive, as fine-mesh discretization produces very large matrices. Recognizing the intrinsic sparsity and low rank of these matrices, we develop a low-rank iterative method (LIM) with stochastic trace estimation that substantially reduces complexity and memory without sacrificing accuracy, while remaining straightforwardly parallelizable. Benchmarks on 2D and 3D topological systems show orders-of-magnitude acceleration over CRM and TMM. LIM extends readily to heterostructures, defective interfaces, and non-Hermitian systems, providing an efficient platform for topological and plasmonic surface phenomena.
Read the full abstract

All authors Kaiqi Liu, Changsheng He, Zihan Xie, Shulin Sun, Lei Zhou, Shaojie Ma

Quantum Physics

Single-photon scattering by a nanofiber-coupled atomic chain: beyond independent radiation

We study single-photon scattering by a nanofiber-coupled atomic chain, comparing a full microscopic model with an independent-radiation approximation. We show that nearly equal transmission probabilities can coexist with substantially different transmitted temporal modes, even after adjusting their relative delay. For a reference chain, a decomposition into collective atomic resonances relates this discrepancy mainly to spectral-phase changes caused by the omitted interatomic couplings. Parameter scans delimit this phase-dominated regime, and controlled comparisons separate the effects of coherent exchange and collective decay.
Read the full abstract

Machine Learning

Beyond the Grid: Continuous Dictionary Pursuit for Interpretable Signal Decomposition

Deep learning models achieve high accuracy in signal processing but remain difficult to interpret. Classical matching pursuit and sparse coding offer better transparency but suffer from spectral leakage and grid mismatch when signal frequencies do not align with discrete dictionary atoms. We introduce Continuous Dictionary Pursuit (CDP), a framework for interpretable signal decomposition that learns the parameters of closed-form functions directly from data. CDP optimizes over a continuous manifold of differentiable atoms and, unlike prior continuous pursuit methods that operate on a single analytic family, searches a heterogeneous library spanning periodic, trend, transient, and discontinuous structures. The algorithm employs a greedy iterative strategy using data-driven priors and gradient-based optimization to isolate individual signal components, with discontinuous waveforms made differentiable through spectral sharpness annealing. To ensure a sparse atom set for reconstruction, the framework utilizes a stopping criterion based on a predefined atom budget and residual error convergence. Evaluations on synthetic and real-world datasets show that CDP removes the grid discretization error for atom families contained in its library, recovers generative parameters to within $1\%$ on stationary synthetic signals expressible in that library, and achieves the best sparsity-accuracy trade-off among classical, continuous-pursuit, and neural baselines in the sparse regime. Beyond decomposition, we demonstrate that the extracted atoms can serve as structured curriculum targets for training sequence models, improving forecasting accuracy over standard end-to-end training on most evaluated datasets. The method offers a glass-box alternative for signal decomposition by providing high reconstruction fidelity alongside direct physical interpretability.
Read the full abstract

Computational Engineering

Rethinking Requirement-to-Code Traceability Through Single-Agent and Multi-Agent Approaches

Requirements-to-code traceability supports essential software engineering activities, including change impact analysis, maintenance, and regulatory compliance. Existing approaches typically rely on a retrieval-and-classification pipeline, which suffers from an inherent retrieval bottleneck: relevant source files that are not retrieved cannot be recovered during subsequent analysis. In this paper, we revisit requirements-to-code traceability in light of the substantially larger context windows offered by modern large language models, which make repository-scale analysis feasible without a separate retrieval step. We propose three approaches: a single-agent approach that analyzes the entire code repository in one pass, and two multi-agent approaches based on self-verification and multi-agent debate. Our evaluation shows that the single-agent approach substantially improves recall, achieving 80.9%, a gain of 38.6 percentage points (pp) over the state-of-the-art retrieval-based baseline, albeit with a 25.8 pp reduction in precision. By leveraging collaborative reasoning, the multi-agent debate-based approach achieves the best overall precision-recall balance, with 24.5% precision and 71.0% recall, yielding an F2 score of 49.5%. This represents an improvement of 7.5 pp over the state of the art and 2.7 pp over the single-agent approach. While the quantitative evaluation establishes the overall effectiveness of the approaches, the aggregate metrics alone explain little due to the considerable variability across results. Importantly, qualitative analysis provides key insights into highly variable quantitative results: the retrieval approach suits requirements that echo the code's vocabulary; single-agent, full-context reasoning suits well-documented, consistently named code; and multi-agent debate suits indirect, conflicting, or distributed evidence.
Read the full abstract

All authors Sarthak Banerjee, Sallam Abualhaija, Domenico Bianculli, Lionel Briand

Software Engineering

GreyZone: Production-quality Agentic Application Threats

Agents are increasingly deployed across domains such as coding, banking, and commerce. Their ability to invoke tools and change application state makes security testing important for reliable deployment. Although agent security has been studied through various adversarial testing approaches, benchmarks grounded in production agent environments remain limited. We introduce GreyZone, a benchmark for evaluating automated red-teaming services against real-world agent applications. Unlike benchmarks such as AgentDojo, which primarily evaluate subject agents under attack, GreyZone evaluates services' ability to generate and deliver attacks that cause verified security-policy violations. We select independently developed, open-source agent applications with security boundaries concerning access permissions, confidentiality, and action authorization. These applications expose diverse attack surfaces, including those arising from multi-agent interactions and multimodal inputs. By grounding evaluation in application environments that preserve production-relevant workflows and independently verifying outcomes, GreyZone supports research on automated agent security testing. Specifically, we present two preliminary case studies using Promptfoo on Letta Code and Clawith to examine attack delivery, agent behavior, and verified outcomes. We release the benchmark, including subject agents and attack surfaces for each application, to support further research on agent testing methods.
Read the full abstract

All authors Ruoyu Zhou, Shuyao Jiang, Haifeng Ruan, Abhik Roychoudhury

Analysis of PDEs

On the optimal mass threshold for the two-dimensional Keller-Segel system coupled to the Navier-Stokes equations

We investigate the mass threshold for the two-dimensional Keller-Segel system coupled with the incompressible Navier-Stokes equations. In the absence of fluid interaction, the classical Keller-Segel system admits a well-known $4\pi$-critical-mass dichotomy: for every suitably regular initial datum of total mass $m<4\pi$, the associated no-flux initial-boundary value problem admits a global bounded classical solution, whereas for $m>4\pi$, there exist smooth initial configurations of mass $m$ such that the associated solution blows up in finite or infinite time. For the fluid-coupled system, Winkler ({SIAM J. Math. Anal.}, 52:2041-2080, 2020) first established global generalized solvability under the condition $m<2\pi$ and posed an open question whether the fluid action may lower the critical blow-up mass to some value lying between $2\pi$ and $4\pi$. Based on numerical evidence, Bonilla and Gutiérrez-Santacreu ({J. Numer. Math.}, 32:175-212, 2024) later conjectured that the critical value is still $4\pi$. In the present work, we resolve this open problem and provide a rigorous mathematical proof to the conjecture, which indicates that the presence of the fluid does not alter the critical mass threshold, and the entire classical subcritical regime remains intact—every admissible initial datum with total mass $m<4\pi$ generates a unique global classical solution that is uniformly bounded in time. Our proof relies on the construction of a new quasi-energy inequality. The key underlying observation is that combining relative isoperimetric estimates with rearrangement techniques yields a Gagliardo-Nirenberg inequality on bounded planar domains with an explicit leading constant.
Read the full abstract

Distributed Computing

UniExit: A Unified Latency-Aware Early-Exit Runtime for Heterogeneous Edge AI

Edge applications increasingly combine CNN-based perception with edge-deployable Transformer-based language models, but deploying such heterogeneous workloads on resource-constrained devices is challenging because inference latency varies with both input difficulty and runtime hardware state. Early-exit inference reduces computation by terminating at intermediate stages, yet existing methods are often model-specific and rely on static latency profiles that become stale under dynamic edge conditions. This paper presents UniExit, a unified latency-aware early-exit runtime for heterogeneous edge AI. UniExit abstracts CNN blocks and Transformer layers as ordered stages and applies a common latency-aware routing policy while retaining model-specific quality mechanisms. It predicts the cost of reaching the next exit from runtime state and inference progress to make budget-aware continue-or-exit decisions. UniExit combines runtime-state-aware latency prediction, cache-based prediction reuse, trend-aware fallback, and CNN-specific exit stabilization to provide low-overhead and robust online control. We implement UniExit on multiple edge and edge-cloud platforms and evaluate it with representative CNN and Transformer-based language-model workloads. Across single-workload, concurrent heterogeneous, and edge-cloud settings, UniExit consistently improves deadline satisfaction and latency-accuracy trade-offs over existing progressive inference baselines, while reducing redundant exit-module computation and task-level energy consumption.
Read the full abstract

All authors Yuhang Guan, Huanle Zhang, Xiaomin Ouyang, Zhidan Liu

Numerical Analysis

A Structured-Surrogate Double-Model Approach to Regularization Parameter Selection

Double-model methods select a regularization parameter by matching the discrepancy between two reconstructions of the same data to a prescribed target. The Structured-Surrogate Double-Model (SSDM) framework is developed around a computationally inexpensive structured surrogate of the forward operator and a spectrally explicit perturbation, with the selected parameter subsequently applied to the original inverse problem. Given a prescribed target, parameter selection requires neither prior knowledge of the noise level nor training data. For a class of Tikhonov regularization problems, the squared discrepancy is shown to attain a strict maximum at zero regularization, decay as $\lambda^{-2}$, and decrease strictly beyond an explicit phase-aware threshold. A complementary, data-dependent certificate is established to guarantee strict monotonicity on finite intervals, enabling unique and reliable target matching even when the global threshold is overly conservative. The discrepancy is further connected to directional roughness in image deblurring, providing a geometric interpretation of the selection principle. Experiments on synthetic images and on synthetically blurred natural photographs illustrate the theoretical findings and demonstrate the practical applicability of certified parameter selection.
Read the full abstract

All authors Malena Español, Markus Juvonen, Samuli Siltanen, Ilaria Trombini, Chuyang Wu

Formal Languages and Automata

Word Subtraction Games

This note is a self-contained text on word subtraction games, combinatorial games in which players alternately remove a nonempty suffix belonging to a prescribed subtraction set from the current word. It collects and develops basic results on these games from the viewpoints of combinatorial game theory and formal language theory. We focus on the relationship between subtraction sets and the resulting languages of winning and losing positions. The topics include finite, regular, and context-free subtraction sets, as well as several variants and further developments.
Read the full abstract

Artificial Intelligence

Evaluation of a Multi-Link Machine Learning Approach for Rainfall Detection Using Satellite Microwave Links

The impacts of climate change have increased the need for reliable real-time rainfall monitoring systems. This study investigates rainfall detection using a network of opportunistic sensors based on satellite microwave links (SMLs) combined with machine learning (ML) classifiers. Two models, a random forest (RF) and a convolutional neural network (CNN), were trained on aggregated SML data, including information from neighboring links. The results demonstrate that leveraging information from multiple links substantially enhances the detection of both wet and dry periods compared to single-link approaches. Moreover, the RF model consistently provides a more stable and reliable trade-off than CNN, improving rainfall detection and reducing misclassifications overall, with particularly strong performance during harsh weather conditions. The study spanned up to 32 rainy days between 2021 and 2024.
Read the full abstract

All authors Mostafa Ftouni, Margherita Spalla, Matteo Colli, Andrea Randazzo, Christian Gianoglio