arXiv preprints from January 1, 2026 through September 27, 2026 — 18:16:28 EST
Decentralized optimization has emerged as a critical paradigm for distributed learning, enabling scalable training while preserving data privacy through peer-to-peer collaboration. However, existing methods often suffer from communication bottlenecks due to frequent synchronization between nodes. We present Overlapping Local...
This work investigates the potential of torrent metadata as a source for open-source intelligence (OSINT), with a focus on user profiling and behavioral analysis. While peer-to-peer (P2P) networks such as BitTorrent are well studied with respect to privacy and performance, their metadata is rarely used for investigative purposes. This...
Context: Quantitative spectroscopy of luminous blue stars relies on detailed non-LTE model atmospheres whose increasing physical realism makes direct, iterative analyses computationally demanding. Aims: We introduce MAUI (Machine-learning Assisted Uncertainty Inference), a statistical framework designed for efficient Bayesian...
With the widespread adoption of large language models (LLMs), hallucinations, which are non-factual fabrications in model outputs, have become serious concerns. Reasoning capabilities have received attention as a self-verification process to improve output reliability. However, the effect of reasoning within a closed system where LLMs...
Rare events in molecular dynamics are often related to noise-induced transitions between different macroscopic states (e.g., in protein folding). A common feature of these rare transitions is that they happen on timescales that are on average exponentially long compared to the characteristic timescale of the system, with waiting time...
Curriculum Learning (CL) aims to improve the outcome of model training by estimating the difficulty of samples and scheduling them accordingly. In NLP, difficulty is commonly approximated using task-agnostic linguistic heuristics or human intuition, implicitly assuming that these signals correlate with what neural models find...
Diffusion inversion is a task of recovering the noise of an image in a diffusion model, which is vital for controllable diffusion image editing. At present, diffusion inversion still remains a challenging task due to the lack of viable supervision signals. Thus, most existing methods resort to approximation-based solutions, which...
In this paper, we consider the geometric setting of navigation data and introduce a natural parallel translation using the Riemannian parallelism. The geometry obtained in this way has some nice and natural features: the natural parallel translation is homogeneous (but in general nonlinear), preserves the Randers type Finslerian norm...
Despite progress in deep learning for Alzheimer's disease (AD) diagnostics, models trained on structural magnetic resonance imaging (sMRI) often do not perform well when applied to new cohorts due to domain shifts from varying scanners, protocols and patient demographics. AD, the primary driver of dementia, manifests through...
Knowledge Distillation (KD) is a central paradigm for transferring knowledge from a large teacher network to a typically smaller student model, often by leveraging soft probabilistic outputs. While KD has shown strong empirical success in numerous applications, its theoretical underpinnings remain only partially understood. In this...
Parallel test-time scaling typically trains separate generation and verification models, incurring high training and inference costs. We propose Advantage Decoupled Preference Optimization (ADPO), a unified reinforcement learning framework that jointly learns answer generation and self-verification within a single policy. ADPO...
Guo and Royle recently classified the connected cubic graphs without eigenvalues of their adjacency matrix in the open interval $(-1, 1)$, and raised the question of extending their classification to graphs of maximum degree at most $3$. They carried out a preliminary investigation of the subcubic case, exhibiting both infinite...
In this paper, we propose a robust real time detection and tracking method for detecting ships in a coastal video sequences. Since coastal scenarios are unpredictable and scenes have dynamic properties it is essential to apply detection methods that are robust to these conditions. This paper presents modified ViBe for moving object...
Large-scale traffic forecasting relies on fixed sensor networks that often exhibit blackouts: contiguous intervals of missing measurements caused by detector or communication failures. These outages are typically handled under a Missing At Random (MAR) assumption, even though blackout events may correlate with unobserved traffic...
NiO/ZnO-based thin films, including single-layer and heterostructure configurations, were synthesized to investigate the influence of stacking order on their electrochemical performance for supercapacitor applications. To improve the relatively low capacitive performance of ZnO compared to NiO, NaCl was introduced as a dopant. All...
We study the impact of background electric fields on a hot plasma of charged particles -- a setting relevant for the early stages of heavy-ion collisions as well as laser pulse experiments. Historically, the electric susceptibility -- encoding the behavior of the hot medium for weak fields -- has been defined within two different...
The anticipated positive social impact of regulatory processes requires both the accuracy and efficiency of their application. Modern artificial intelligence technologies, including natural language processing and machine-assisted reasoning, hold great promise for addressing this challenge. We present a framework to address the...
We present a derivation using kinetic wave theory of the two-dimensional empirical Garrett--Munk spectrum for ocean internal waves, valid at all frequencies including near-inertial frequencies. This is based directly on the governing equations for a two-dimensional Boussinesq system with constant stratification and rotation. Our...
Recently, diffusion models have achieved a great performance with a small dataset of size $n$ and a fast optimization process. However, the estimation error of diffusion models suffers from the curse of dimensionality $n^{-1/D}$ with the data dimension $D$. Since images are usually a union of low-dimensional manifolds, current works...
We recently characterized the separated determinantal point processes $Λ_φ$ associated with Fock spaces $\mathcal F_φ$ in the plane with doubling weight $φ$. We also showed that, as expected, a more restrictive condition is required to characterize the separated Poisson processes with the same first intensities as $Λ_φ$. To gain...
Graph neural networks (GNNs) have achieved breakthroughs in various real-world downstream tasks due to their powerful expressiveness. As the scale of real-world graphs has been continuously growing, a storage-based approach to GNN training has been studied, which leverages external storage (e.g., NVMe SSDs) to handle such web-scale...
Tape diagrams provide a graphical notation for categories equipped with two monoidal products, $\otimes$ and $\oplus$, where $\oplus$ is a biproduct. Recently, they have been generalised to handle Kleisli categories of arbitrary monoidal monads. In this work, we show that for the subdistribution monad, tapes are isomorphic to...
Estimating causal effects of continuous treatments is a common problem in practice, for example, in studying average dose-response functions. Classical analyses typically assume that all confounders are fully observed, whereas in real-world applications, unmeasured confounding often persists. In this article, we propose a novel...
We report on the modal analysis of open gyrotropic waveguides (GWs). The GWs consist of a non-circular gyrotropic (i.e., gyroelectric and gyromagnetic) core and an infinitely extending isotropic cladding. To solve this problem, we develop two independent full-wave methods. The first technique is an extended integral equation (EIE)...
We investigate a set of design principles that link specific features of interparticle interactions to predictable structural and dynamic outcomes in two-dimensional self-assembly, a framework relevant to soft matter and biological condensates. Using extensive Molecular Dynamics simulations of single- and two-component systems, we...