arXiv preprints from January 1, 2026 through September 23, 2026 — 11:34:24 EST
Recent advances in Reinforcement Learning with Verifiable Rewards (RLVR) for Large Language Model (LLM) reasoning have been hindered by a persistent challenge: exploration collapse. The semantic homogeneity of random rollouts often traps models in narrow, over-optimized behaviors. While existing methods leverage policy entropy to...
For a $C^1_{t,x}$ solution $u$ to the incompressible 3D Euler equations, the helicity $H(u(t))=\int_{\mathbb{T}^3} u \cdot \textrm{curl}\, u$ is constant in time. For general low-regularity weak solutions, it is not always clear how to define the helicity, or whether it must be constant in time in the case that there is a clear...
Sequential Bayesian optimal experimental design (SBOED) for PDE-governed inverse problems is computationally challenging, especially for infinite-dimensional random field parameters. High-fidelity approaches require repeated forward and adjoint PDE solves inside nested Bayesian inversion and design loops. We formulate SBOED as a...
Ultra-dense crowds, in which physical contact between people cannot be avoided, pose major safety concerns. Nevertheless, the underlying dynamics driving their collective behaviours remain poorly understood. Existing dense crowd models, mostly two-dimensional and contact-based, overlook biomechanical mechanisms that govern individual...
Retrieval-Augmented Generation (RAG) models are critically undermined by citation hallucinations, a deceptive failure where a model cites a source that fails to support its claim. While existing work attributes hallucination to a simple over-reliance on parametric knowledge, we reframe this failure as an evolving, scale-dependent...
We introduce the first method for change-point detection on encrypted time series. Our approach employs the CKKS homomorphic encryption scheme to detect shifts in statistical properties (e.g., mean, variance, frequency) without ever decrypting the data. Unlike solutions based on differential privacy, which degrade accuracy through...
With the growth of interpreting technologies, from remote interpreting and Computer-Aided Interpreting to automated speech translation and interpreting avatars, there is now a high demand for ways to quickly and efficiently measure the quality of any interpreting delivered. A range of approaches to fulfil the need for quick and...
We prove a version of the Poincaré-Bendixson theorem for certain classes of curves on the 2-sphere which are not required to be the trajectories of an underlying flow or semiflow on the sphere itself. Using this result we extend the Poincaré-Bendixson theorem to the context of continuous semiflows on compact subsets of the 2-sphere...
We consider an optimal control problem governed by a rate-inde\-pendent system with non-convex energy. The state equation is approximated by means of viscous regularization w.r.t.\ to hierarchy of two different Hilbert spaces. The regularized problem corresponds to an optimal control problem subject to a non-smooth ODE in Hilbert...
We introduce Ising-Hüsler-Reiss processes, a new class of multivariate Lévy processes that allows for sparse modeling of the path-wise conditional independence structure between marginal stable processes with different stability indices. The underlying conditional independence graph is encoded as zeroes in a suitable precision matrix....
Among the many recommendations of the Decadal Survey on Astronomy and Astrophysics 2020, we found that a priority area of research is to pave the pathways towards finding and characterizing habitable worlds. In this context, we aim to understand how planetary systems evolve through atmospheric escape, and develop techniques to...
Community detection is a fundamental task in data analysis, and block models provide an approach for identifying a wide variety of community structures while offering high interpretability. The degree-corrected block model (DCBM) is an established model that accounts for the heterogeneity of node degrees. However, inference methods...
Given any countable group $G$, we construct uncountably many quasi-isometry classes of proper geodesic metric spaces with quasi-isometry group isomorphic to $G$. Moreover, if the group $G$ is a hyperbolic group, the spaces we construct are hyperbolic metric spaces. We make use of a rigidity phenomenon for quasi-isometries exhibited...
The objective of this work is to assess computationally efficient coarse-grained plasticity models against high-fidelity crystal plasticity simulations for magnesium polycrystals over a wide range of textures and grain sizes. A basic requirement is that such models are able to capture {\it evolving} plastic anisotropy and...
In this study, we explore temperature-dependent CPT violation during Big Bang Nucleosynthesis (BBN) through electron-positron mass asymmetries parametrized by $b_0(T) = αT^2$. The $T^2$ scaling naturally evades stringent laboratory bounds at zero temperature while allowing for significant CPT violation at MeV scales in the early...
Baryonic processes such as radiative cooling and feedback from massive stars and active galactic nuclei (AGN) directly redistribute baryons in the Universe but also indirectly redistribute dark matter due to changes in the gravitational potential. In this work, we investigate this "back-reaction" of baryons on dark matter using...
Deep cognitive attention is characterized by heightened gamma oscillations and coordinated visual behavior. Despite the physiological importance of these mechanisms, computational studies rarely synthesize these modalities or identify the neural regions most responsible for sustained focus. To address this gap, this work introduces...
We analyze the mixing time of open quantum systems governed by the Lindblad master equation, showing that it is determined not only by the Liouvillian gap, but also by the trace-norm factor of each decaying Liouvillian eigenmode. By utilizing them as universal predictors of mixing time, we establish general conditions for the fast and...
We present a search for and characterization of ionized outflows in 15 star-forming systems at $z\sim3-9$ with no evidence of Active Galactic Nuclei (AGN), observed with JWST/NIRSpec IFU as part of the GA-NIFS program. The targets often show satellites and complex substructure, from which we isolate 40 galaxies/regions. The sample...
We study the sensitivity of past and future stopped-pion experiments to light, invisible dark sector particles produced in exotic pion decays. We consider two-body decays involving sterile neutrinos, $π^+ \to \ell^+ N$, as well as three-body decays $π^+ \to \ell^+ ν_\ell X$, with $X$ an invisible scalar, axion-like particle, or dark...
We present a novel and powerful constraint on the physics of supermassive black hole (BH) accretion disks. We show that in the outer disk (radii $R \gtrsim 0.01\,$pc or $\gtrsim 1000\,R_{G}$), models supported by thermal or radiation pressure predict disk masses which are much larger than the BH mass and increase with radius - i.e....
Massive stars are often found in multiple systems, yet how binary-star systems with very close separations ($\lesssim$ au) assemble remains unresolved. We investigate the formation and inward migration of massive-star binaries in Solar-metallicity environments using the star-cluster formation simulation of Chon et al. (2024), which...
We study the one-loop renormalization of the Georgi-Machacek model. At one loop, the renormalization of the model is phenomenologically important when triggered by operators that are absent at the tree level due to the global $SU(2)_R$ symmetry. By computing all the tree-level parameters from the standard input parameters $α_e$,...
We realize bosonic string theory with ensembles of infinite width neural networks. The string tension is tuned by the variance of the output weights. The construction provides a new computation of the foundational Virasoro-Shapiro and Veneziano amplitudes as neural network correlators.
Disentangling turbulence and bulk motions in the intracluster medium (ICM) of galaxy clusters is inherently ambiguous, as the plasma is continuously stirred by different processes on disparate scales. This poses a serious problem in the interpretation of both observations and numerical simulations. In this paper, we use filtering...