arXiv preprints from January 1, 2026 through September 25, 2026 — 04:53:37 EST
Vision-Language Models (VLMs) have shown remarkable performance in User Interface (UI) grounding tasks, driven by their ability to process increasingly high-resolution screenshots. However, screenshots are tokenized into thousands of visual tokens (e.g., about 4700 for 2K resolution), incurring significant computational overhead and...
Index tracking, also known as passive investing, has gained significant traction in financial markets due to its cost-effective and efficient approach to replicating the performance of a specific market index. This review paper provides a comprehensive overview of the various modeling approaches and strategies developed for index...
The deployment of Large Vision-Language Models (LVLMs) for real-world document question answering is often constrained by dynamic, user-defined policies that dictate information disclosure based on context. While ensuring adherence to these explicit constraints is critical, existing safety research primarily focuses on implicit social...
We identify a polarization rotation systematic in the far field beams of refractive cosmic microwave background (CMB) telescopes caused by differential transmission in anti-reflection (AR) coatings of optical elements. This systematic was identified following the development of a hybrid physical optics method that incorporates...
Image deblurring is a challenging problem in imaging due to its highly ill-posed nature. Deep learning models have shown great success in tackling this problem but the quest for the best image quality has brought their computational complexity up, making them impractical on anything but powerful servers. Meanwhile, recent works have...
Scaling superconducting quantum computers to the fault-tolerant regime calls for a commensurate scaling of the classical control and readout stack. Today's systems largely rely on room-temperature, rack-based instrumentation connected to dilution-refrigerator cryostats through many coaxial cables. Looking ahead, superconducting...
Sybil attacks remain a fundamental obstacle in open online systems, where adversaries can cheaply create and sustain large numbers of fake identities. Existing defenses, including CAPTCHAs and one-time proof-of-personhood mechanisms, primarily address identity creation and provide limited protection against long-term, large-scale...
Optical polarimetry provides information on the geometry of the emitting region, the magnetic field configuration and the properties of dust in astrophysical sources. Current state-of-the-art instruments typically have a small field of view (FoV), which poses a challenge for conducting wide surveys. We propose the construction of the...
This paper addresses the nonparametric estimation of a spatially varying, heteroskedastic variance function on the unit sphere within a regression framework. While adaptive regression estimation is well-established on manifolds, characterizing localized noise structures presents unique theoretical obstacles due to bias propagation...
We study when geometric simplicity of decision boundaries, used here as a notion of interpretability, can conflict with accurate approximation of axis-aligned decision trees by shallow neural networks. Decision trees induce rule-based, axis-aligned decision regions (finite unions of boxes), whereas shallow ReLU networks are typically...
We study the expansions of permutation statistics in the basis of functions counting occurrences of a fixed pattern in a permutation. We show the finiteness of these pattern expansions for a class of permutation statistics including the higher moment statistics, generalizing a result of Berman and Tenner. We also give a combinatorial...
The Time-Slotted Channel Hopping (TSCH) mode of IEEE802.15.4 standard provides ultra high end-to-end reliability and low-power consumption for application in field of Industrial Internet of Things (IIoT). With the evolving of Industrial 4.0, dynamic and bursty tasks with varied Quality of Service (QoS); effective management and...
Although the measurement of current is now defined with respect to the electronic charge, producing a current standard based on a single-electron source remains challenging. The error rate of a source must be below 0.01 ppm, and many such sources must be operated in parallel to provide practically useful values of current in the...
Microscopic evaluation of white blood cell morphology is central to leukemia diagnosis, yet current deep learning models often act as black boxes, limiting clinical trust and adoption. We introduce HemBLIP, a vision language model designed to generate interpretable, morphology aware descriptions of peripheral blood cells. Using a...
Presentaremos una nueva demostración del teorema de Shafarevich sobre finitud de curvas elípticas con buena reducción fuera de un conjunto finito de primos dado. Esto da un nuevo punto de entrada a teoremas fundamentales de finitud diofantina tales como el teorema de Siegel sobre la ecuación $S$-unidad. Nuestro argumento está libre de...
Building on insights from the grokking literature, we study character-level Transformers trained to compute modular addition from text, and focus on robustness under input-format variation rather than only in-distribution accuracy. We identify a previously under-emphasized failure mode: models that achieve high in-distribution...
Multiple-choice question answering (MCQA) is easy to evaluate but adds a meta-task: models must both solve the problem and output the symbol that *represents* the answer, conflating reasoning errors with symbol-binding failures. We study how language models implement MCQA internally using representational analyses (PCA, linear probes)...
A nearly half-cubic hectometer of deep Mediterranean-Sea waters is yearlong sampled with about 3000 high-resolution temperature sensors to study different sources of turbulent waterflows, which are vital for life. Although temperature differences are never larger than 0.01degrC, daily, weekly, and seasonal variations are observed....
Semileptonic $τ$ decays mainly proceed via interactions between charged lepton and quark currents. The hadronization of the quark current is intrinsically nonperturbative and generally cannot be addressed analytically. In these proceedings, we propose using symmetry arguments alone to construct clean angular observables, which, within...
We study the mutual singularity of countable Gaussian mixture models (GMMs), with particular emphasis on infinite-dimensional settings. We first establish that a countable mixture of Gaussian probability measures is itself a well-defined probability measure. We then prove a general measure-theoretic result showing that if every...
Geometric and Topological Deep Learning are rapidly growing research areas that enhance machine learning through the use of geometric and topological structures. Within this framework, Group Equivariant Non-Expansive Operators (GENEOs) have emerged as a powerful class of operators for encoding symmetries and designing efficient,...
The asymptotic distribution of the likelihood-ratio statistic for testing parameters on the boundary is well known to be a chi-squared mixture. The mixture weights have been shown to correspond to the intrinsic volumes of an associated tangent cone, unifying a wide range of previously isolated special cases. While the weights are...
Retrieval-augmented generation (RAG) enhances large language models (LLMs) by incorporating external knowledge, but existing approaches indiscriminately trigger retrieval and rely on single-path evidence construction, often introducing noise and limiting performance gains. In this work, we propose Decide Then Retrieve (DTR), a...
This paper presents a neuromorphic, event-driven tactile sensing system for soft, large-area skin, based on the Dynamic Vision Sensors (DVS) integrated with a flexible silicone optical waveguide skin. Instead of repetitively scanning embedded photoreceivers, this design uses a stereo vision setup comprising two DVS cameras looking...
Many nonlinear optimal control and optimization problems involve constraints that combine continuous dynamics with discrete logic conditions. Standard approaches typically rely on mixed-integer programming, which introduces scalability challenges and requires specialized solvers. This paper presents an exact reformulation of broad...