arXiv preprints from January 1, 2026 through September 23, 2026 — 04:50:29 EST
This paper addresses the minimization of a finite sum of prox-convex functions under Lipschitz continuity of each component. We propose two variants of the splitting proximal point algorithms proposed in \cite{Bacak,Bertsekas}: one deterministic with a fixed update order, and one stochastic with random sampling, and we extend them...
Transformer channel decoders, such as the Error Correction Code Transformer (ECCT), have shown strong empirical performance in channel decoding, yet their generalization behavior remains theoretically unclear. This paper studies the generalization performance of ECCT from a learning-theoretic perspective. By establishing a connection...
Using deep medium-resolution JWST rest-optical spectra of a sample of typical star-forming galaxies (Lyman break galaxies and Lyman-$α$ emitters) from the LyC22 survey at $z \sim 3$, we determined the nebular abundances of N, O, and Ne relative to H for a subsample of 25 objects with the direct method, based on auroral [OIII]4363 line...
The Newton method has been widely adopted to achieve certified unlearning. A critical assumption in existing approaches is that the data requested for unlearning are selected i.i.d.(independent and identically distributed). However,the problem of certified unlearning under non-i.i.d. deletions remains largely unexplored. In practice,...
As Large Language Models (LLMs) evolve from static dialogue interfaces to autonomous general agents, effective memory is paramount to ensuring long-term consistency. However, existing benchmarks primarily focus on casual conversation or task-oriented dialogue, failing to capture **"long-term project-oriented"** interactions where...
Unified large multimodal models (LMMs) have achieved remarkable progress in general-purpose multimodal understanding and generation. However, they still operate under a ``one-size-fits-all'' paradigm and struggle to model user-specific concepts (e.g., generate a photo of \texttt{<maeve>}) in a consistent and controllable manner....
Wake interactions in floating wind farms are inherently coupled to platform motion, yet most experimental studies to date neglect this two-way coupling by prescribing platform movements. This work presents a hardware-in-the-loop (HIL) wind-tunnel methodology to investigate wake interactions between two floating wind turbines with...
Type-II hyperbolic lattices constitute a new class of hyperbolic structures that are projected onto the Poincaré ring and possess both an inner and an outer boundary. In this work, we reveal the higher-order topological phases in type-II hyperbolic lattices, characterized by the generalized quadrupole moment. Unlike the type-I...
Neutron star asteroseismology offers a unique opportunity to probe nuclear physics through stellar oscillations. Although the pressure inside of a neutron star is typically assumed to be isotropic, pressure anisotropy can arise from various physical mechanisms, including elasticity, viscosity, and magnetic fields. Previous studies of...
The success of deep neural networks largely depends on the statistical structure of the training data. While learning dynamics and generalization on isotropic data are well-established, the impact of pronounced anisotropy on these crucial aspects is not yet fully understood. We examine the impact of data anisotropy, represented by a...
A recently proposed generalised continuum theory of curved dislocations describes the spatial and temporal evolution of statistically stored and geometrically necessary dislocation densities as well as the curvature. The dynamics follow from a scalar plastic potential that constrains the allowed velocity fields and leads to a phase...
Post-training quantization is essential for deploying Large Language Models (LLMs) on resource-constrained devices. However, standard integer quantization (e.g., INT4) fundamentally degrades performance by imposing a uniform grid on the heavy-tailed distribution of weight parameters, particularly in smaller-scale models (e.g., <2B...
We report the presence of a flavor-violating correction to the $VAAA$-type anomaly, $δ^{(\prime)}$, induced by the surface terms of the anomalous quark triangle diagram, previously found in the $η,η'\toπ^+π^-γ$ decay amplitudes, and investigate its impact on the corresponding semileptonic decay modes of $η$ and $η'$. The magnitude of...
In this paper, we give a new proof to a past stability result established in Fournodavlos-Rodnianski-Speck (arXiv:2012.05888), for Kasner solutions of the $(3+1)$-dimensional Einstein vacuum equations under polarized $U(1)$-symmetry. Our method, inspired by Beyer-Oliynyk-Olvera-Santamar{\'ı}a-Zheng (arXiv:1907.04071,...
We study IndIGO-D, a decihertz gravitational-wave mission concept, focusing on a specific configuration in which three spacecraft fly in formation to form an L-shaped interferometer in a heliocentric orbit. The two orthogonal arms share a common vertex, providing a space-based analogue of terrestrial Michelson detectors, while...
Flow structures directly responsible for local skin-friction generation in turbulent channel flows are identified using the newly developed Canonical Correlation Decomposition (CCD) method. The dominant structures take the form of streamwise streaks that are spanwise-localised around the position where the skin-friction is targeted...
The classical Dirichlet problem on the unit disk can be solved by different numerical approaches. The two most common and popular approaches are the integration of the associated Poisson integral and, by applying Dirichlet's principle, solving a particular minimization problem. For practical use, these procedures need to be...
Competitive programming poses a significant challenge for Code LLMs. While recent models have shown promise, they heavily rely on finite real-world data, raising concerns about scalability and contamination. In this paper, we investigate a critical question: Can we elevate models to expert-level reasoning performance using fully...
It is shown that the familiar two-stream instability is the result of spontaneous parity-time (PT)-symmetry breaking in a conservative system, and more importantly, explicit PT-symmetry breaking by viscosity can destabilize the system in certain parameter regimes that are stable when viscosity vanishes. This reveals that complex...
We confront the ghost-free non-local $F(R)$ gravity theories with the latest Atacama Cosmology Telescope (ACT) constraints on the spectral index of the scalar perturbations and the updated constraints of Planck/BICEP on the tensor-to-scalar ratio. After reviewing how the ghost-free non-local version of $F(R)$ gravity can be obtained,...
Let $(Ω^{n+1}, g)$ be an $(n+1)$-dimensional smooth compact connected Riemannian manifold with smooth boundary $Σ$, satisfying that ${\text{Ric}_Ω}\ge 0$ and $Σ$ is strictly convex, more precisely, its second fundamental form $h\ge cg_Σ$ for some positive constant $c$. Escobar {\cite{escobar1997geometry}} considered the first nonzero...
Let $A$ be a unital locally matrix algebra. Among the examples of such algebras are: (1) an infinite tensor product $\otimes M_{n_i}(\mathbb{F})$ of matrix algebras over a field $\mathbb{F}$, and (2) the Clifford algebra of a nondegenerate quadratic form on an infinite-dimensional vector space over an algebraically closed field of...
Two $F(R)$ gravity models are tested on the basis of their viability during all stages of cosmological evolution. It is shown that these models can describe both the early-time inflationary epoch and the dark energy epoch. The models are confronted with the latest observational data, including the Pantheon+ catalogue with Type Ia...
Open-source software (OSS) is a critical component of modern software systems, yet supply chain security remains challenging in practice due to unavailable or obfuscated source code. Consequently, security teams often rely on runtime observations collected from sandboxed executions to investigate suspicious third-party components. We...
The benefits of most large language models come with steep and often hidden economic and environmental costs due to their resource usage inefficiency during deployment. Model quantization improves energy and memory efficiency through representing model parameters by lower-precision values. However, compression below 4-bits often...