arXiv preprints from January 1, 2026 through September 13, 2026 — 06:35:36 EST
We propose Ambient Dataloops, an iterative framework for refining datasets that makes it easier for diffusion models to learn the underlying data distribution. Modern datasets contain samples of highly varying quality, and training directly on such heterogeneous data often yields suboptimal models. We propose a dataset-model...
VIBETENSOR is an open-source research system software stack for deep learning, generated by LLM-powered coding agents under high-level human guidance. In this paper, "fully generated" refers to code provenance: implementation changes were produced and applied as agent-proposed diffs; validation relied on agent-run builds, tests, and...
Sparse-view Cone-Beam Computed Tomography reconstruction from limited X-ray projections remains a challenging problem in medical imaging due to the inherent undersampling of fine-grained anatomical details, which correspond to high-frequency components. Conventional CNN-based methods often struggle to recover these fine structures, as...
Inspired by regular black holes (RBHs) sourced by dark matter halos, we generalize the anisotropic energy-momentum tensor by relaxing the $P_r = -ρ$ condition between radial pressure and density. We demonstrate that while RBHs are a unique special case, a broader class of relations yields horizonless compact stars. Under specific...
Identifying and understanding non-BCS superconductivity remains a central challenge in condensed-matter physics. Here we focus on the Sc$_{6}$MTe$_{2}$ family (M =d-electron metal), which provides a unique platform of isostructural compounds exhibiting superconductivity across 3d, 4d, and 5d systems. Using hydrostatic pressure as an...
Establishing ages for young clusters is key for properly tracking the star formation history of a region. In this paper we investigate a new approach to estimating ages for young populations, based on the well-founded assumption that the initial mass function is the same throughout a star forming cloud. We trial this method for six...
Mental health concerns are rising globally, prompting increased reliance on technology to address the demand-supply gap in mental health services. In particular, mental health chatbots are emerging as a promising solution, but these remain largely untested, raising concerns about safety and potential harms. In this paper, we dive into...
We examine a class of stochastic differential inclusions involving multiscale effects designed to solve a class of generalized variational inequalities. This class of problems contains constrained convex non-smooth optimization problems, constrained saddle-point problems and various equilibrium problems in economics and engineering....
We study the existence of a decomposition of the diagonal for bidegree hypersurfaces in a product of projective spaces. Using a cycle theoretic degeneration technique due to Lange, Pavic and Schreieder, we develop an inductive procedure that allows one to raise the degree and dimension starting from the quadric surface bundle of...
This is a very short introduction to hierarchically hyperbolic spaces and groups. It is aimed at non-experts, including anyone who may encounter a group with some similarities to mapping class groups.
Medical vision-language models can automate the generation of radiology reports but struggle with accurate visual grounding and factual consistency. Existing models often misalign textual findings with visual evidence, leading to unreliable or weakly grounded predictions. We present CURE, an error-aware curriculum learning framework...
We consider mixed four-point correlators of 1/2-BPS operators $\mathcal{O}_{k}$ in the maximally supersymmetric CFTs, i.e. the 3d $\mathcal{N}=8$, 4d $\mathcal{N}=4$, and 6d $\mathcal{N}=(2,0)$ theories. In arXiv:hep-th/0405180, Dolan, Gallot, and Sokatchev demonstrated that four-point correlators of identical $\mathcal{O}_{k}$ in...
Multimodal Large Language Models (MLLMs) have recently demonstrated strong performance on a wide range of vision-language tasks, raising interest in their potential use for biometric applications. In this paper, we conduct a systematic evaluation of state-of-the-art MLLMs for heterogeneous face recognition (HFR), where enrollment and...
We extend to multiplicative lattices a theorem of Anderson and Roitman characterizing the cancellation ideals of a commutative ring.
We introduce a new family of polar-like codes, called Partially Polarized Polar (PPP) codes. PPP codes are constructed from conventional polar codes by selectively pruning polarization kernels, thereby modifying the synthesized bit-channel capacities to ensure a guaranteed number of non-frozen bits available early in decoding. These...
In 2012, K. Matsuda introduced the class of weakly closed graphs and investigated when binomial edge ideals are F-pure. He proved that weakly closed binomial edge ideals are F-pure whenever the base field has positive characteristic. He conjectured that: (i) when the base field has characteristic two, every F-pure binomial edge ideal...
We describe $H^p(V)$, a subset of $p$-rough path space $Ω_p(V)$ which is a vector space under an addition operation $\boxplus$ and a scalar multiplication $\odot$. We show that the domain of $\boxplus$ can be extended to $Ω_p(V)\times H^p(V)$, allowing any $p$-rough path $X$ to be additively perturbed by an $H\in H^p(V)$. We prove...
Homomorphic encryption (HE) enables arithmetic operations to be performed directly on encrypted data. It is essential for privacy-preserving applications such as machine learning, medical diagnosis, and financial data analysis. In popular HE schemes, ciphertext multiplication is only defined for two inputs. However, the multiplication...
Accreting systems can launch powerful outflows which interact with the surrounding medium. We combine new radio observations of the accreting neutron star X-ray binary (XRB) Circinus X-1 (Cir X-1) with archival radio observations going back 24 years. The $\sim3$ pc scale wide-angle radio and X-ray emitting caps found around Cir X-1...
High-Performance Computing (HPC) schedulers must balance user performance with facility-wide resource constraints. The task boils down to selecting the optimal number of nodes for a given job. We present a surrogate-assisted multi-objective Bayesian optimization (MOBO) framework to automate this complex decision. Our core hypothesis...
Beck and Teboulle's FISTA for finding the minimizer of the sum of two convex functions is one of the most important algorithms of the past decades. While function value convergence of the iterates was known, the actual convergence of the iterates remained elusive until October 2025 when Jang and Ryu, as well as Boţ, Fadili, and Nguyen...
The rapid emergence of new entities -- driven by cultural shifts, evolving trends, and personalized user data -- poses a significant challenge for existing Speech Large Language Models (Speech LLMs). While these models excel at general conversational tasks, their static training knowledge limits their ability to recognize...
Certain families of quantum mechanical models can be described and solved efficiently on a classical computer, including qubit or qudit Clifford circuits and stabilizer codes, free-boson or free-fermion models, and certain rotor and GKP codes. We show that all of these families can be described as instances of the same algebraic...
User interactions with language models vary due to static properties of the user (trait) and the specific context of the interaction (state). However, existing persona datasets (like PersonaChat, PANDORA etc.) capture only trait, and ignore the impact of state. We introduce Chameleon, a dataset of 5,001 contextual psychological...
Knowledge Distillation (KD) is increasingly adopted to transfer capabilities from large language models to smaller ones, offering significant improvements in efficiency and utility while often surpassing standard fine-tuning. Beyond performance, KD is also explored as a privacy-preserving mechanism to mitigate the risk of training...