arXiv preprints from January 1, 2026 through September 21, 2026 — 08:41:54 EST
To stabilize the training of Large Language Models (LLMs), gradient clipping is a nearly ubiquitous heuristic used to alleviate exploding gradients. However, traditional global norm clipping erroneously presupposes gradient homogeneity across different functional modules, leading to an adverse "spill-over" effect where volatile...
Retrieval-Augmented Generation systems depend on retrieving semantically relevant document chunks to support accurate, grounded outputs from large language models. In structured and repetitive corpora such as regulatory filings, chunk similarity alone often fails to distinguish between documents with overlapping language....
We extend the Rate-Distortion-Perception (RDP) framework to the Rényi information-theoretic regime, utilizing Sibson's $α$-mutual information to characterize the fundamental limits under distortion and perception constraints. For scalar Gaussian sources, we derive closed-form expressions for the Rényi RDP function, showing that the...
Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, represents a major global pandemic of the 21st century, with long-term effects termed long COVID. This systematic review and network meta-analysis (NMA) evaluated pharmacological interventions for adults with long COVID, incorporating randomized controlled trials and adjusted...
Clinical AI systems frequently suffer performance decay post-deployment due to temporal data shifts, such as evolving populations, diagnostic coding updates (e.g., ICD-9 to ICD-10), and systemic shocks like the COVID-19 pandemic. Addressing this ``aging'' effect via frequent retraining is often impractical due to computational costs...
The evolution toward 6G networks increasingly relies on network slicing to provide tailored, End-to-End (E2E) logical networks over shared physical infrastructures. A critical challenge is effectively decomposing E2E Service Level Agreements (SLAs) into domain-specific SLAs, which current solutions handle through computationally...
DEAP-3600 is a single-phase liquid argon (LAr) direct-detection dark matter experiment operating 2 km underground at SNOLAB (Sudbury, Canada). The detector consists of 3.3 tons of LAr contained in a spherical acrylic vessel. At a WIMP mass of 100 GeV, DEAP-3600 has a projected sensitivity of $10^{-46}\,\mathrm{cm}^{2}$ for the...
Neuromorphic computing offers a pathway toward energy-efficient processing of data, yet hardware platforms combining nanoscale integration and multimodal functionality remain scarce. Here we demonstrate a gallium-phosphide electro-optomechanical spiking neuron that integrates optical and electromechanical interfaces within a single...
We extend the CIPSI-driven CC($P$;$Q$) methodology [K. Gururangan et al., J. Chem. Phys. 155 (2021) 174114], in which the leading higher-than-doubly excited determinants are identified using the selected configuration interaction (CI) approach abbreviated as CIPSI, to excited electronic states via the equation-of-motion (EOM)...
Our purpose in this paper is to construct new examples of twisted Brill Noether loci on curves of genus g greater than 2 with negative expected dimension. We begin by completing the proof of Butler's conjecture for coherent systems of certain type establishing the birationality, smoothness, and irreducibility of the corresponding...
Recent advances in task-oriented dialogue (TOD) systems, driven by large language models (LLMs) with extensive API and tool integration, have enabled conversational agents to coordinate interleaved goals, maintain long-horizon context, and act proactively through asynchronous execution. These capabilities extend beyond traditional TOD...
As the only known unambiguous star in a Maunder minimum-like chromospheric activity state, the properties of HD 166620 can provide valuable insight into the behaviour of the Sun during the historic extended low-states of its activity cycle. The coronal X-ray activity of HD 166620 has so far only been probed with a ROSAT/HRI...
We establish the global existence of a class of weak solutions to the isentropic compressible Navier-Stokes equations in a half-plane with Dirichlet boundary conditions, allowing for vacuum both in the interior and at infinity, under a suitably small initial total energy. The solutions constructed here admit unbounded densities and...
Standard fractional models on manifolds often conflate geometric anisotropy with medium heterogeneity. In this Letter, we overcome this rigidity by deriving the fundamental solution for a weighted space-time fractional ultrahyperbolic operator, denoted by $(-\Box_{φ,ω})^β$. Using a novel spectral approach based on the Weighted Fourier...
The increasing use of generative artificial intelligence (GenAI) in qualitative research raises important questions about analytic practice and interpretive authority. This study examines how researchers interact with an Inductive Thematic Analysis GPT (ITA-GPT), a purpose-built AI tool designed to support inductive thematic analysis...
The influence of oxygen stoichiometry on the uranium local environment is explored in epitaxial single crystal uranium oxide thin films grown by DC magnetron sputtering. Through post-growth annealing, the stoichiometry of as-grown UO$_{2}$ films are tuned over an approximate stoichiometry range of $0.07 \leq x \leq 0.20$, estimated...
As AI systems grow increasingly capable of operating for hours or days at a time, users' prompts are transforming into elaborate specifications for the AI to autonomously work on. While prompting for bounded, single-turn tasks has been extensively studied, less is known about how people communicate specifications for long-horizon...
In a digraph $D$,a quasi-kernel is an independent set $Q$ such that for every vertex $u$, there is a vertex $v \in Q$ satisfying $\text{dist}(v,u)\leq 2$. In 1974 Chvátal and Lovász showed every digraph contains a quasi-kernel. In 1976, P. L. Erdős and Székely conjectured that every sourceless digraph has a quasi-kernel of order at...
We present results and analyses from the third VoicePrivacy Challenge held in 2024, which focuses on advancing voice anonymization technologies. The task was to develop a voice anonymization system for speech data that conceals a speaker's voice identity while preserving linguistic content and emotional state. We provide a systematic...
Recent developments in causal machine learning methods have made it easier to estimate flexible relationships between confounders, treatments and outcomes, making unconfoundedness assumptions in causal analysis more palatable. How successful are these approaches in recovering ground truth baselines? In this paper we analyze a new data...
The paper presents an improved upper bound (achievability result) on the optimal tradeoff between Normalized Delivery Time (NDT) and computation load for distributed computing MapReduce systems in certain ranges of the parameters. The upper bound is based on interference alignment combined with zero-forcing. The paper further provides...
Patterning of precipitates along dislocation lines arising from nonequilibrium segregation during ion irradiation is investigated in model binary alloys. Lattice kinetic Monte Carlo simulations reveal that the competition between solute advection by point defects to the dislocation and thermal diffusion along the dislocation can...
Gravitational-wave (GW) observatories have used template-based search to detect hundreds of compact binary coalescences (CBCs). However, template-based search cannot detect astrophysical sources that lack accurate waveform models, including core-collapse supernovae, neutron star glitches, and cosmic strings. Here, we present a novel...
Dominating Set is a well-known combinatorial optimization problem which finds application in computational biology or mobile communication. Because of its $\mathrm{NP}$-hardness, one often turns to heuristics for good solutions. Many such heuristics have been empirically tested and perform rather well. However, it is not well...
Large Language Models (LLMs) have shown strong performance on code understanding tasks, yet they fundamentally lack the ability to perform precise, exhaustive mathematical reasoning about program behavior. Existing benchmarks either focus on mathematical proof automation, largely disconnected from real-world software, or on...