arXiv preprints from January 1, 2026 through September 25, 2026 — 16:51:02 EST
Human cognition integrates information across nested timescales. While the cortex exhibits hierarchical Temporal Receptive Windows (TRWs), local circuits often display heterogeneous time constants. To reconcile this, we trained biologically constrained deep networks, based on scale-invariant hippocampal time cells, on a language...
Reservoir computing is a bio-inspired machine learning paradigm that exploits the intrinsic dynamics of nonlinear systems with fading memory for efficient temporal information processing. Microelectromechanical resonators offer a promising platform for reservoir computing as they inherently possess the requisite nonlinear and temporal...
The time-discretized, spatially continuous generalized Euler equations are a prototype example of multi-marginal optimal transport, yet the question whether they exhibit mass-splitting (or equivalently, whether they have solutions that are not of Monge form) has remained open. Here we resolve this question by giving a mass-splitting...
For $b>0$ and $n\geqslant 1$, we consider the $n\times n$ tridiagonal matrix $V_n(b)$ with diagonal entries $b+1$, superdiagonal entries $1$, and subdiagonal entries $b$. A diagonal similarity reduces $V_n(b)$ to a symmetric tridiagonal Toeplitz matrix and hence makes its spectrum explicit. Since $\det\left(V_n(b)\right)$ equals the...
We present GKFieldFlow, a novel three-dimensional autoregressive deep learning surrogate model for nonlinear gyrokinetic turbulence. Based on the architecture FieldFlow-Net, this model combines a multi-resolution 3D U-Net encoder-decoder that operates on evolving plasma potential fields. A dilated temporal convolutional network (TCN)...
The rapid advancement of wireless communication technologies, including 5G, emerging 6G networks, and the large-scale deployment of the Internet of Things (IoT), has intensified the need for efficient spectrum utilization. Automatic modulation classification (AMC) plays a vital role in cognitive radio systems by enabling real-time...
We show that, under certain constraints, the Stanley-Reisner ring of an infinite simplicial complex is Cohen-Macaulay in the sense of ideals and weak Bourbaki unmixed. We apply this result to prove the wanted claim -- that initial complexes of matrix Schubert varieties corresponding to infinite permutations in $S_{\infty}$ with...
By 2025, holotomography (HT) has matured from a niche optical modality into a versatile platform for quantitative, label-free imaging in biomedicine. By reconstructing the three-dimensional refractive-index (RI) distribution of cells and tissues, HT enables high-resolution volumetric imaging with low phototoxicity and minimal sample...
We introduce AfriEconQA, a specialized benchmark dataset for African economic analysis grounded in a comprehensive corpus of 236 World Bank reports. The task of AfriEconQA is to answer complex economic queries that require high-precision numerical reasoning and temporal disambiguation from specialized institutional documents. The...
Conformal novelty detection is a classical machine learning task for which uncertainty quantification is essential for providing reliable results. Recent work has shown that the BH procedure applied to conformal p-values controls the false discovery rate (FDR). Unfortunately, the BH procedure can lead to over-optimistic assessments...
Large language model fine-tuning is bottlenecked by memory: a 7B parameter model requires 84GB--14GB for weights, 14GB for gradients, and 56GB for FP32 optimizer states--exceeding even A100-40GB capacity. We present Chronicals, an open-source training framework achieving 3.51x speedup over Unsloth through four synergistic...
We study a class of maps having the Collatz function (famously related to the Collatz Conjecture) as an example, under topological and ergodic perspectives, including an approach with thermodynamic formalism. By introducing a key topology and its Borel sigma-algebra we show that recurrence implies periodicity. Moreover, we establish...
We present high-resolution near-infrared spectroscopic observations of the newly identified supernova remnant (SNR) candidate G25.8+0.2 obtained with the Immersion Grating Infrared Spectrograph (IGRINS) on the Harlan J. Smith Telescope. The source was originally discovered in the UKIRT Wide-field Infrared Survey for Fe+ (UWIFE; Lee et...
Monocular depth estimation aims to recover the depth information of 3D scenes from 2D images. Recent work has made significant progress, but its reliance on large-scale datasets and complex decoders has limited its efficiency and generalization ability. In this paper, we propose a lightweight and data-centric framework for zero-shot...
Some deep learning-based point cloud registration methods struggle with zero-shot generalization, often requiring dataset-specific hyperparameter tuning or retraining for new environments. We identify three critical limitations: (a) fixed user-defined parameters (e.g., voxel size, search radius) that fail to generalize across varying...
Photonic integrated circuits require robust room-temperature single-photon sources to enable scalable quantum technologies. Single-walled carbon nanotubes (CNTs), with their unique excitonic properties and chemical tunability, are attractive candidates, but their integration into photonic circuits remains challenging. In this work, we...
Existing change detection methods often lack the versatility to handle diverse real-world queries and the intelligence for comprehensive analysis. This paper presents a general agent framework, integrating Large Language Models (LLM) with vision foundation models to form ChangeGPT. A hierarchical structure is employed to mitigate...
We investigate an intrinsic step-jamming phenomenon at the nanometer scale on Kardar-Parisi-Zhang (KPZ)-like kinetically roughened crystal surfaces that arises during interface-limited steady crystal growth or retreat. Monte Carlo simulations using the Metropolis algorithm on a restricted solid-on-solid (RSOS) lattice model...
We obtain infinitely many boundary operators in the Brownian loop soup in the subcritical phase by analyzing the conformal block expansion of the two-point function that computes the probability of having two marked points on the upper half-plane being separated by Brownian loops. The resulting boundary operators are primary operators...
With the rapid development of e-commerce, auto-bidding has become a key asset in optimizing advertising performance under diverse advertiser environments. The current approaches focus on reinforcement learning (RL) and generative models. These efforts imitate offline historical behaviors by utilizing a complex structure with expensive...
This paper discusses the task of face-based speech synthesis, a kind of personalized speech synthesis where the synthesized voices are constrained to perceptually match with a reference face image. Due to the lack of TTS-quality audio-visual corpora, previous approaches suffer from either low synthesis quality or domain mismatch...
Large Language Model (LLM)-based agents are increasingly deployed in e-commerce applications to assist customer services in tasks such as product inquiries, recommendations, and order management. Existing benchmarks primarily evaluate whether these agents successfully complete the final task, overlooking the intermediate reasoning...
Most membership inference attacks (MIAs) against Large Language Models (LLMs) rely on global signals, like average loss, to identify training data. This approach, however, dilutes the subtle, localized signals of memorization, reducing attack effectiveness. We challenge this global-averaging paradigm, positing that membership signals...
Early detection of fake news is critical for mitigating its rapid dissemination on social media, which can severely undermine public trust and social stability. Recent advancements show that incorporating propagation dynamics can significantly enhance detection performance compared to previous content-only approaches. However, this...
The evolution of Large Language Models (LLMs) from passive text generators to autonomous, goal-driven systems represents a fundamental shift in artificial intelligence. This chapter examines the emergence of agentic AI systems that integrate planning, memory, tool use, and iterative reasoning to operate autonomously in complex...