Programming Languages
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 14 September
Managed language runtimes often provide concurrent garbage collectors so that latency-critical applications with large working sets can keep running while most collection work proceeds in the background. ZGC is a production-quality, generational, concurrent collector in OpenJDK with sub-millisecond pause times. While ZGC is designed to run concurrently, frequent and excessive collections with ZGC can still slow the mutators due to synchronization costs and interference with shared computing resources. Hence, the ZGC scheduler is conservative by default, and in most cases, will grow the heap toward the maximum allowed before scheduling a collection. While this approach minimizes collection effort, it can be wasteful, or even harmful, if the maximum heap size is not well tuned to the actual working set. We propose Opportunistic ZGC (OppZGC), a feedback-directed ZGC scheduling policy that constrains the heap dynamically and automatically, without per-application tuning. OppZGC identifies periods when CPU cores are underutilized and leverages them for concurrent collection with ZGC. We describe the design and implementation of OppZGC in OpenJDK's HotSpot Java VM and evaluate it with standard and latency-sensitive benchmarks from DaCapo Chopin and SPECjbb. OppZGC limits heap usage when there is CPU capacity sufficient for additional collections, and avoids scheduling extra collections when they would substantially degrade performance. Overall, it reduces maximum heap usage for our DaCapo benchmarks by between 61% and 90%, on average, depending on configuration, with minimal impact on throughput and request latency compared to default ZGC.
Classical Physics
Increase of the electromechanical coupling of piezoelectric vibration harvesters through lateral bars
To enhance the performance of vibration energy harvesters, it is essential to maximise the electromechanical coupling coefficient k${}^2$ of piezoelectric devices. This enables sufficient harvested power and tuning capability of the resonant frequency through electrical methods. While much literature treats the optimisation of piezoelectric cantilevers, the optimisation range is usually limited by the transverse coupling coefficient k31${}^2$ of the material. This work introduces an innovative solution to extend the optimisation range and increase the coupling coefficient of piezoelectric cantilevers. This is achieved by minimising lateral strain in the beam using lateral bars to maximise the equivalent material coupling coefficient. The theoretical basis of this innovation is demonstrated through the exploitation of the constitutive equations of piezoelectric materials. The interest of the addition of lateral bars to increase the coupling coefficient is demonstrated and studied with simulations based on the finite elements method. Finally, a proof of concept is realised using a aluminum cantilever prototype integrating a lead-free KNaNbO3 (KNN) piezoelectric material. It is tested under vibration at 0.1 m/s${}^2$ with variable resistive loads. The results show that the coupling coefficient increases by 30% and the relative frequency bandwidth by 32% with resistive tuning of the resonant frequency, by fixing eight steel bars to the cantilever. The designed prototype is a highly performant leadfree vibration energy harvester. It produces a maximum power of 49.9 $\mu$W at resonance, and its normalised power density is equal to 16.5 mW/G${}^2$/cm${}^3$. Its relative frequency bandwidth is equal to 3.1%.
Read the full abstract
Functional Analysis
A sharp bound for sampling widths in the uniform norm: kernel $D$-optimal designs and oversampling
For every reproducing kernel Hilbert space $\mathcal{H}_K$ with bounded kernel $K$, we prove that the linear sampling widths $g_m^{\text{lin}}$ and Gelfand widths $c_n$ in the uniform norm satisfy $$ g_m^{\text{lin}}(B_{\mathcal H_K})_\infty \leq \frac{m+1}{m-n+1}\, c_n(B_{\mathcal H_K})_\infty\quad , \quad m\ge n. $$ In a certain sense, the leading factor $(m+1)/(m-n+1)$ on the right-hand side is sharp. We transfer the problem into the selection of a maximal volume subset of kernel translates (kernel $D$-optimal design). Apart from basic linear algebra, our proof employs low rank approximation techniques and a version of the Eckart-Young-Mirsky theorem. The same argument gives corresponding bounds for the selection of a reduced basis in Hilbert spaces.
Read the full abstract
Machine Learning
Bayesian Optimisation Using Product-of-Experts Gaussian Process Models with Uncertainty Calibration
Bayesian optimisation (BO) typically relies on a single global Gaussian process (GP) model as its surrogate model. However, GP regression has cubic computational complexity in the number of training data points, limiting its applicability to large-scale optimisation problems. The product-of-experts Gaussian process model with uncertainty calibration (GP-pro-c) mitigates this limitation by combining multiple local GP experts, enabling improved uncertainty quantification, reduced computational cost, and preservation of global correlations. Despite these desirable properties, the use of GP-pro-c in BO has not been thoroughly studied. This paper introduces BO-pro-c, a Bayesian optimisation algorithm that uses GP-pro-c as its surrogate model, and evaluate its performance across a diverse range of BO settings. Experimental results suggest that BO-pro-c maintains competitive optimisation performance while achieving a 0.9% reduction in simple regret and a 39.4% reduction in computational overhead relative to a BO algorithm based on a single global GP model.
Read the full abstract
Metric Geometry
Greedy Packing of Nested Rings: Placement Rules, a Golden Counterexample, and a Tribonacci Floor
We study packings of annuli of a common width, allowing each ring to nest inside the hole of a larger one. The objectives of maximizing contact area and cardinality diverge: area is superadditive in the radius, cardinality is not. Under superincreasing radii, every descending greedy maximizes every positive, strictly increasing, superadditive objective. More strongly, any choice among feasible containers yields the lexicographically maximal feasible set, for containers of arbitrary shape in every dimension. This placement irrelevance holds unconditionally for at most three rings and fails at four in disks and squares; twin instances exclude every universal rule based only on the observable state. Write $\rho=\max_i(\sum_{j>i}r_j)/r_i$. The additive model has threshold exactly $1$. For disks we prove the exact global threshold $\tau=\varphi$, with no failure at $\rho\le\varphi$, for every finite inventory, even with independent hole radii. The key geometric theorem states that, under golden tail bounds, an entire disk list fits a circular container if and only if its three largest disks fit; this supplies the uniform exchange of parents that the threshold proof needs. The Tribonacci constant $T\approx1.83929$ remains the exact floor of a rigid subfamily. A dimension-reduction lemma transfers spherical sharpness results to all dimensions $d\ge2$, and a separate argument proves the golden threshold for at most five rings in those dimensions. For square pans, a Cartesian confinement criterion gives twins and a family proving $\tau_{\square}\le Y\approx1.6845$; its optimality is open. For independent holes, the exact universal area guarantee under $\rho\le\kappa<1$ is $\min(1,\kappa^{-2}-1)$, with threshold $1/\sqrt2$. The repository has 122 Lean theorems. Euclidean geometry, forest assembly and continuity remain written proofs; numerical checks do not substitute for them.
Read the full abstract
Probability
Phase transition in optimal hypercontractivity
We discover an exponent-dependent phase transition phenomenon for optimal hypercontractivity: for every prescribed $q_0>2$, there exists a reversible continuous-time Markov chain on three state space with normalized spectral gap whose $(2,q)$-optimal hypercontractivity time satisfies $$ \text{$t_{\mathrm{opt}}(2,q)=\frac12\log(q-1)$ if and only if $q\ge q_0$},$$ whereas the strict inequality $t_{\mathrm{opt}}(2,q)>\frac12\log(q-1)$ holds for $2<q<q_0$.
Read the full abstract
Biomolecules
Co-folding with a Soup of Representations
Co-folding models such as AlphaFold3, Protenix, ESMFold2, and OpenDDE have advanced rapidly, yet no single model consistently performs best across all biomolecular complexes. In this paper, we show that their pair representations encode complementary information that can be transferred across models to improve structure prediction. We introduce SoupFold, which combines pair representations from multiple co-folding models in a common representation space and generates structures from the combined representation. Importantly, SoupFold does not retrain the co-folding models and learns only simple mappings to transfer representations across models. We evaluate SoupFold on antibody-antigen, protein-protein, protein-ligand, molecular glue, GPCR, and oligomeric complex prediction using AlphaFold3, Protenix, ESMFold2, and OpenDDE. By combining representations across models, SoupFold improves over individual co-folding models across the considered benchmarks.
Read the full abstract
Materials Science
Gate-tunable magnetic damping in van der Waals Heterostructures
Active control of magnetic damping is a crucial capability for the development of low-power, tunable spintronic devices. In this work, we demonstrate that the intrinsic magnetic damping of 2D materials is highly sensitive to the position of the Fermi level relative to spin-orbit driven anti-crossings, or “spin hot spots.” Starting from a minimal multi-band model evaluated via Kamberský's breathing Fermi surface theory, we elucidate the fundamental role that energy gaps and band alignment play in governing magnetic dissipation. We then translate these theoretical insights to van der Waals heterostructures, proposing a concrete mechanism for strongly tunable magnetic damping via electrostatic gating. This concept is validated from first principles via density functional theory calculations on a realistic Fe$_3$GeTe$_2$/graphene heterostructure. Our calculations reveal that an out-of-plane electric field selectively shifts the relevant energy bands, enabling the modulation of the heterostructure's magnetic damping over several orders of magnitude. These findings establish a robust theoretical framework and a promising materials platform for electrically tunable magnetization dynamics.
Read the full abstract
Computational Geometry
Multi-Stage NeRF for Efficient 3D Coronary Artery Reconstruction from Two Narrow-Angle Angiographic Projections
X-ray coronary angiography is the clinical gold standard for coronary artery disease during real-time cardiac interventions, but provides only 2D projections of inherently 3D vessels. Existing learning-based 2D-to-3D reconstruction methods typically require wide angular coverage or multiple views, assumptions that are rarely satisfied in routine practice where only two projections with narrow angular separation are available. To address these challenges, we propose NeCA++, a multi-stage self-supervised neural radiance field (NeRF) framework tailored to clinically realistic acquisition constraints. The framework decomposes reconstruction into two stages that progressively refine spatial support and representation capacity. In the first stage, a coarse 3D representation of the vasculature is reconstructed, restricting the subsequent optimisation to regions with a higher likelihood of vessel presence, termed an active region. Afterward reconstruction is restricted to this region while higher-resolution representations are progressively activated to recover fine vascular details. This multi-stage strategy focuses learning on anatomically plausible regions, mitigates gradient dilution under extreme sparsity, and stabilises global topology before recovering fine vascular branches. Furthermore, two vessel-specific regularisations are introduced: a ray-aligned constraint to reduce projection-induced ambiguity, and a bimodal density penalty to enable early vessel-background separation. Extensive experiments across three datasets (ImageCAS, ASOCA, and Synthetic RCA) and four angular configurations demonstrate consistent superiority over state-of-the-art baselines, particularly under clinically realistic narrow-angle settings, while achieving reconstruction within 58 seconds per case.
Read the full abstract
Geometric Topology
On possible values of the signature of flat unitary bundles over compact surfaces
We determine all possible signatures of flat Hermitian bundles over compact, connected, oriented surfaces of positive genus with nonempty boundary. If $\Sigma_{g,n}$ has genus $g\geq 1$ and $n\geq 1$ boundary components, then the signatures arising from representations $\pi_1(\Sigma_{g,n})\to\mathrm{U}(p,q)$ are exactly the integers $m$ satisfying \[ |m| \leq (p+q)(2g+n-2) -(2g-2)|p-q| -\min\{2,n|p-q|\}. \] Every such integer, including the two extremal values, is realized by a block-diagonal representation whose image is contained, after possibly interchanging $p$ and $q$, in $\mathrm{U}(1,1)^{\times\min\{p,q\}}\times\mathrm{U}(|p-q|)$.
Read the full abstract
Human-Computer Interaction
Who Chooses the Artwork? Curatorial Agency and Distributed Intent in Botto
Botto is often described as a decentralized autonomous artist, but its authorship cannot be located in image generation alone. This paper examines Botto as an agentic curatorial system in which generation, ranking, voting, feedback, and minting form a recursive loop. Drawing on Botto's documentation and prior accounts of the project, we argue that agency in Botto is distributed but asymmetrical: community participants influence artistic direction through voting and governance, while Botto's internal models structure which images become visible, selectable, and recognizable as works. The case shows that, for AI art systems, explainability should include not only model behavior but also the curatorial procedures through which outputs acquire artistic status.
Read the full abstract
Astrophysics of Galaxies
The Multiphase CGM in the Epoch of Reionization: CII and CIV absorbers around [OIII] Emitters
We investigate the multiphase circumgalactic medium (CGM) during the Epoch of Reionization (EoR, $z>6$) by cross-correlating cool- CII and warm-ionized CIV absorption systems with star-forming [OIII] emitters. JWST/NIRCam wide-field slitless spectroscopy from the EIGER survey is combined with medium- and high-resolution optical/NIR spectra of six background quasars, including new VLT/X-Shooter observations of PSO J159-02. We analyze the relation between 16 CII and 14 CIV absorbers ($\log(N) > 13.0$) and 136 galaxies, within an impact parameter of R$_{\perp} \leq$ 1000 pkpc and a line-of-sight separation of $\Delta v \leq 500$ km/s. We detect a statistically significant excess of both ions around galaxies compared to a randomized background. We observe that the CIV covering fraction remains enhanced up to $\sim 1$ pMpc, whereas CII drops to the background level beyond $\sim 0.5$ pMpc, demonstrating that the warm-ionized phase is more spatially extended than cooler gas. Jointly, the 3D galaxy-absorber spatial clustering is significantly weaker than the galaxy-galaxy auto-correlation. This provides direct physical evidence that early carbon enrichment is not confined to the virial radius of massive star-forming systems; rather, a substantial fraction of these metals permeates the diffuse intergalactic medium (IGM) or is injected by a widespread population of faint, undetected dwarf galaxies. Finally, we note a rapid radial decline of the CIV covering fraction compared with lower redshift samples at $z<2$ and $z\sim3-4$ pointing out an evolving CGM ionization structure where early metals reside predominantly in lower ionization states. In conclusion, we determine a conservative lower limit for the observed carbon mass of $\rm M_{CII + CIV} \geq 2.8 \times 10^6 \, M_{\odot}$ within 300 pkpc.
Read the full abstract
Classical Physics
Sliding contact fraction in gravity-driven dense cohesionless granular flows
Using large-scale, three-dimensional discrete element simulations, we investigate the kinematic properties of granular chute flows of cohesionless grains, together with the evolution of the coordination number and the fraction of sliding contacts. Our results reveal that, under certain conditions, increasing the grain–grain friction coefficient leads to a decrease in the global dissipation of the system. This finding highlights that the overall dissipation is not solely governed by the energy lost during sliding contacts, but also by the probability that such contacts occur, a probability that decreases markedly as the friction coefficient increases. Together, these effects demonstrate the subtle and nontrivial interplay between grain-scale frictional interactions and macroscopic flow behavior in dense and cohesionless granular materials.
Read the full abstract
Machine Learning
The Token Before the Value Is the Key
Induction circuits retrieve a value by matching a query to information carried from the token before that value. How this address forms, incorporates context, and supports sequential retrieval remains less clear. We use hybrid language models as a tractable testbed and identify the historical source key as a common causal interface between local preparation and global retrieval. Probes and restoration interventions connect this interface across controlled models, pretrained recurrent and sliding-window hybrids, and released language models. We organize the analysis by address context radius and sequential retrieval depth. First, shared local parameters couple predecessor Carrying to recurrent state processing; separating these demands or biasing initialization promotes address formation. Second, adapting input representations extends an established literal circuit to category matching, making semantically compatible records active competitors. Earlier context resolves this competition through the source key, with a sharp visibility boundary in a controlled single-read model and compositional relays across local layers. Third, fixed relations support memorization shortcuts, whereas sequential generalization requires learning an output-to-query bridge between retrievals. Controlled state-tracking and compositional in-context tasks demonstrate this requirement across stacked stages, looped recurrence, and intermediate-token generation. Together, these results connect circuit formation, contextual disambiguation, and sequential use through a shared retrieval interface. Code is available at https://github.com/ckpassenger/bind-match-copy.
Read the full abstract
Machine Learning
Specifying Reward Functions for RL Without Environment Sampling
Enabling human stakeholders to specify reward functions that lead to their desired outcomes is a key challenge in deploying reinforcement learning agents. Preference-based methods such as online RLHF can reduce the burden of manual reward design, but they require repeatedly training policies, sampling trajectories from the real world, and eliciting feedback, making them impractical in settings where environment interaction is computationally expensive or unsafe. We introduce Experience-Free Autonomous Reward Specification (EARS), a method for learning reward functions from preferences without environment interaction. Our approach uses a structured LLM-mediated process to construct a small set of expressive reward features from a task description and the environment observation space, then strategically samples imagined trajectories in this feature space and learns feature weights from preferences over the imagined trajectory pairs. We evaluate on three long-horizon domains: pandemic lockdown regulation design, insulin administration for diabetes patients, and autonomous vehicle control on a highway. We compare EARS to baselines that also enable reward specification without environment interaction–namely, methods that directly prompt an LLM to generate a reward function. When learning from either ground-truth preference labels or preferences labeled by a LLM, EARS designs reward functions that are more aligned with the ground truth reward function that produced the preferences or LLM context than these baselines. These results suggest that preference-based reward specification remains effective without environment sampling, enabling practical reward design in settings where collecting real trajectories is costly or infeasible.
Read the full abstract
Dynamical Systems
A mathematical modelling of rheumatoid arthritis
We developed a mathematical model to describe the inflammatory dynamics during arthritis development and resolution, focusing on the interaction between three key immune cell populations within the joint tissue, namely neutrophils (N), sublining synovial tissue-resident macrophages (SL-TRM, T), and circulating monocytes that are recruited into the inflamed joint and differentiate into monocyte-derived macrophages (MoMac, M). An inflammatory signal (S), representing the response to an external arthritogenic stimulus, drives the recruitment and activation of neutrophils and MoMac, while SL-TRM exert a regulatory and protective role. We propose a minimal ordinary differential equation model of acute inflammatory response into a joint that couples an inflammatory signal S to the three cell populations (N, T, M) and incorporates recruitment, stimulation, nonlinear self-limitation and mutual regulation. We show that the equilibrium structure depends on three adimensional parameters A, B, C and derive a cubic equation characterizing nontrivial steady states. We perform an analytical study of equilibria and their stability, and identify a bifurcation at A = 1 separating a disease-free regime from a sustained inflammatory state. Numerical simulations illustrate typical transient dynamics under various scenarios and show the ability of the model to capture the transition from acute to chronic inflammation, as well as the resolution of inflammation under certain conditions. In conclusion, our model provides a framework to explore mechanisms underlying acute versus sustained joint inflammation and offers a basis for parameter estimation from experimental data, as well as future extensions toward evaluation of therapeutic strategies.
Read the full abstract
Cosmology and Nongalactic Astrophysics
Anisotropic redshift distributions in photometric galaxy clustering and their cosmological impact
Photometric galaxy clustering is a major probe of large-scale structure whose interpretation relies on accurate characterization of the galaxy redshift distribution, $n(z)$. Standard analyses assume the $n(z)$ of a selected lens sample to be isotropic, whereas observational systematics can imprint spatial variations. We quantify how this anisotropy biases angular galaxy clustering, internal consistency, and cosmological inference. We develop an analytical formalism for clustering under spatially varying $n(z)$ and forward-model realistic observational systematics, galaxy selection, and tomographic binning. For a fiducial magnitude-limited sample, modeling $w(\theta)$ with a single global $\bar n(z)$ underestimates the clustering signal by up to $\sim10\%$ at small scales in the highest-redshift bin, decreasing towards lower redshift. In a $2\times2$pt analysis, cosmological parameter shifts remain below $1\sigma$ for KiDS-like and DES-like setups. For an LSST-like setup retaining KiDS-like selection and observing-condition variation, $\sigma_8$ is biased low by $1.7\sigma$ and $\Omega_m$ high by $1.4\sigma$, while the high-redshift galaxy bias shifts by up to $\sim4\sigma$. The anisotropy also affects whether the probes are consistent enough to combine: a calibrated posterior-predictive internal-consistency test finds the contaminated angular clustering improbable given the joint cosmic-shear and galaxy–galaxy-lensing constraints (median calibrated $\tilde p\approx8\times10^{-4}$ under LSST-like noise). Anisotropic redshift distributions are therefore a potentially significant systematic for future photometric clustering analyses. Our forward model can be used to mitigate it given a realistic selection function and galaxy population model.
Read the full abstract
Physics and Society
The life and death of football team runs: survival of collective modes shapes Lévy-like transport
Broad run-length distributions and short-lag superdiffusion have recently been reported in football players and team centroids, prompting a collective-foraging interpretation. The mechanism linking these player- and team-level signatures remains unclear. Using SoccerMon GPS data from 66 tracked team-match records across 62 fixtures in the Norwegian women's top-flight Toppserien, we asked whether player transport is inherited from a translating team mode and what controls the lifetime of that mode. We decomposed player displacement into centroid translation and motion within the formation, then modelled the switching and termination of directional centroid runs. At longer lags, centroid translation carried an increasing share of player displacement. Run lifetimes were broad but finite: termination was highest near onset, declined with age and rose modestly later. After accounting for speed and directional persistence, collective order added only limited predictive information. At the population level, an age-dependent switching-and-termination model predicted, on held-out match dates, the early enrichment and later depletion of high-order states among surviving runs. These results link player- and team-level Lévy-like movement through a finite-lived collective mode whose age and internal state shape its survival. Long displacements need not be selected in advance; they can emerge when transient collective modes persist, reorganise and selectively survive. More broadly, trajectory statistics may record which dynamical histories survive, not only the rules by which agents move.
Read the full abstract
General Relativity and Quantum Cosmology
Higher-Dimensional Slowly Rotating Black Holes in Nonlinear Electrodynamics
We study slowly rotating, nonlinearly charged black holes in arbitrary spacetime dimension $d \geq 4$, using a generalized Lense–Thirring ansatz. Working to linear order in the rotation parameters, we solve the Einstein–nonlinear electrodynamics (NLE) field equations for restricted theories ${\cal L}({\cal S})$ depending on single invariant. We provide an extension of the four-dimensional "No Go theorem" proven in ArXiv:2203.01919 and demonstrate the uniqueness of Maxwell theory in $d \geq 4$. Similar to four dimensions, it is proven that the standard rotation generating trick fails to produce full rotating solutions for any nontrivial NLE in higher dimensions. We identify a family of NLE theories admitting slowly rotating solutions with vector potentials encoding rotation effect in Maxwell-like manner; in four dimensions this reproduces the previously found RegMax theory, while in $d\geq4$ we obtain the corresponding electric potential in closed form and the associated Lagrangian implicitly as a function of the radial coordinate.
Read the full abstract
Data Structures and Algorithms
Strong and Compact Policies for Submodular Markov Decision Processes via LP-Based Submodular Orienteering
Finding policies for Markov Decision Processes (MDPs) is a central problem in areas such as Reinforcement Learning and Operations Research. Here, we have to repeatedly choose an action that should be performed by an agent. Depending on the action and the current state of the agent, the agent collects a reward and randomly transitions into a new state. The goal is to maximize the reward in expectation over a finite time horizon of length $H$. We consider a recently introduced variant that generalizes the traditionally additive reward function in the model to a monotone submodular one, which allows for capturing a range of interesting applications. Without the stochastic component, this problem is equivalent to the Submodular Orienteering problem, where the goal is to find an $s$-$t$ walk in a directed graph maximizing a monotone submodular function under a length constraint. We present a novel LP-based algorithm for Submodular Orienteering using ideas from the Sherali-Adams hierarchy and Round-or-Cut. Our guarantees are comparable to the known quasi-polynomial time logarithmic approximation for Submodular Orienteering, but also extend to the setting of Submodular Markov Decision Processes. In the polynomial time regime, we present an $O(n^{\varepsilon})$-approximation (and $O(H^{\varepsilon})$ for Submodular MDPs) for every $\varepsilon >0$, where $n$ is the number of vertices, which was unknown even for Submodular Orienteering. Prior to our work, the best known approximation guarantee for Submodular MDPs had an approximation ratio linear in $H$. Beyond these algorithmic results, our methods reveal a trade-off between the approximation guarantee and the number of previously visited vertices on which an agent conditions its decision.
Read the full abstract
Probability
Critical branching random walks on $\mathbb{Z}^2$: local survival probabilities and Yaglom limit theorems
We consider a branching random walk on $\mathbb{Z}^2$ with critical offspring of mean $1$ and spatial motion governed by the jumps of a lazy simple random walk. For every site $x\in\mathbb{Z}^2$, we obtain a uniform asymptotical estimate for the local survival probability, i.e., the probability that there are particles at $x$ at large time $n$. Using Stein's method, we establish a Yaglom-type theorem for the number of particles at $x$ at time $n$ when $x$ is at distance of order $\sqrt{n}$ from the origin. Moreover, at the position occupied by a typical particle at time $n$, the number of particles at that site, normalized by $\log n$, converges in law to a Gamma distribution, thereby confirming a conjecture of Lalley and Zheng [Ann. Probab. 39 (2010), 327-368]. Finally, we prove that, conditional on local survival, the total number of particles at time $n$, divided by $n$, also converges weakly to a Gamma distribution.
Read the full abstract
Systems and Control
Safe Newton-Based Extremum Seeking for Static Maps with Delayed Output Measurements
This work presents a delayed safe Newton-based extremum seeking (SANES) framework for minimizing an unknown static map subject to an unknown safety constraint. The objective and safety measurements are assumed to be affected by the same constant time delay. To compensate for delayed measurements, a model-free predictor is developed to construct the quantities required for optimization, including the nominal Newton-based extremum-seeking control input and the gradient information used to formulate control Lyapunov function (CLF) and control barrier function (CBF) conditions. Robust CLF–CBF quadratic programs (QPs), subject to parameter-update constraints, are then formulated to account explicitly for derivative-estimation and prediction errors. For the delay-free case, robustness margins are derived from bounds on the extremum-seeking estimation errors, whereas for the delayed case, the margins incorporate both estimation and prediction errors. Practical stability of the nominal Newton-based extremum-seeking dynamics is established directly through a Lyapunov analysis, thereby providing convergence guarantees over the admissible parameter set without relying exclusively on local averaging arguments. Robust CLF and CBF conditions are subsequently derived to establish practical convergence and forward invariance of a robust subset of the prescribed safe set. A numerical case study demonstrates the effectiveness of the proposed SANES framework in achieving constrained optimization despite unknown objective and safety maps and delayed measurements.
Read the full abstract
Numerical Analysis
Physics-Guided Conditional Flow Matching with Energy Regularization for Robust PDE Inverse Problems
We consider partial differential equation (PDE) inverse problems from sparse, noisy, and corrupted observations, with the aim of recovering unknown coefficient fields and associated state variables in a mesh-free setting. Under such sparse and corrupted observations, standard physics-informed and generative approaches typically treat all samples indiscriminately and therefore lack a principled mechanism for reconciling physical laws with contaminated data. We address this difficulty with a two-stage flow-matching framework. In the first stage, we develop physics-guided conditional flow matching (PG-CFM), which incorporates strong-form PDE information through residual regularization along the generative trajectories together with intermittent global collocation constraints. In the second stage, we introduce energy-regularized flow matching (ERFM), which fine-tunes the Stage-1 model by assigning each observation a physics–data energy score from a frozen teacher and reweighting the flow-matching objective to reduce the influence of high-energy, PDE-inconsistent samples. We show that the resulting Stage-2 objective is equivalent to flow matching under a teacher-induced reweighted data distribution, which gives a population-level interpretation of the robustness mechanism. Numerical experiments on several inverse benchmarks, including Poisson and Navier–Stokes problems, show that the proposed framework yields more accurate coefficient recovery than robust PINN variants and competing generative baselines.
Read the full abstract
Astrophysics of Galaxies
ALMA-JELLY II. Constraining the radial profiles of the quenching timescales in ram-pressure-stripped galaxies
We investigate the environmental quenching histories and spatial progression of ram-pressure stripping in 11 galaxies in the nearby Coma, Norma, and A1367 clusters. As part of the ALMA-JELLY large program, we combine high-sensitivity VLT/MUSE integral-field spectroscopy with ultraviolet-to-far-infrared photometry in 37 spatially resolved disc apertures. The sample spans $\log(M_*/\mathrm{M}\odot)\sim7.5-10.5$ and includes galaxies undergoing active or recent ram-pressure stripping, identified through prominent H$\alpha$ or UV tails. We use Bayesian spectrophotometric modelling with exponentially declining star-formation histories and show with mock data that combining spectra and photometry breaks parameter degeneracies and improves quenching-age constraints. All galaxies show quenched outer discs and relatively old ($\lesssim1$ Gyr) stellar populations. Massive galaxies ($\log(M_*/\mathrm{M}_\odot)>9.5$) exhibit clear outside-in quenching: inner regions quench about 500 Myr later than their outskirts. The small scatter in quenching age with distance from the truncation radius, despite differences in environment and stripping geometry, points to a common evolutionary sequence. Low-mass and dwarf galaxies instead show weak radial gradients, with quenching occurring more rapidly, typically within 300 Myr.
Read the full abstract
Analysis of PDEs
Harnack estimates for nondivergence kinetic equations
We establish a weak Harnack inequality for nonnegative strong supersolutions and a Harnack inequality for nonnegative strong solutions of the nondivergence-form kinetic equation \[ \partial_tu+v\cdot\nabla_xu+a(t,x,v):\nabla_v^2u +b(t,x,v)\cdot\nabla_vu+c(t,x,v)u=0, \] where the coefficients $a$, $b$, and $c$ are merely Borel measurable, the matrix $a$ is uniformly elliptic, and $b$ and $c$ are bounded. As a consequence, we derive interior kinetic Hölder estimates for strong solutions. Our approach develops a probabilistic kinetic analogue of the Krylov–Safonov method. The main new ingredient is a quantitative version of Krylov's estimate for kinetic Itô processes with progressively measurable coefficients, which in turn yields the global existence of weak solutions to the associated kinetic SDE with bounded measurable coefficients. These results appear to be the first Harnack and Hölder regularity theory for kinetic equations with merely measurable coefficients.
Read the full abstract
Machine Learning
The Misery of Mechanistic Interpretability: A Formal Perspective
Mechanistic interpretability has become the dominant lens for understanding frontier language models, as their inner workings are complex and inherently black boxes. To gain insights into these models, interpretable replacement networks (IRNs) are trained at all layers, exposing interpretable features through sparsely activated neurons. However, the faithfulness of an IRN is usually evaluated only empirically on clean data, and we show that even semantically minor input perturbations flip the dominant IRN features-and thus the human-understandable interpretation-across five open-weight model families (GPT-2 small, Gemma 2 2B, Gemma 3 1B, Llama 3.2 1B, R1-Distill-Qwen 1.5B). We propose the first formal verification framework for the faithfulness of an IRN, where reachability analysis certifies a sound upper bound of the faithfulness gap in adversarial scenarios. Moreover, we show that verification-aware training of IRNs substantially tightens this certified bound, restoring a feature-level interpretation that safety auditors can act on. Together, these results give, to the best of our knowledge, the first formal guarantees for mechanistic interpretability of large language models.
Read the full abstract
Probability
Stability in stochastic hypergraph matching III: general reneging
In many real-life matching problems, waiting agents might abandon before being matched, such as patients deceasing before receiving organs, passengers/drivers cancelling ride requests, or raw materials/intermediary products degrading in production lines. This poses the need for incorporating reneging in stochastic matching models. In this work, we consider matching models on hypergraphs with batch arrivals and general-weight matchings. Since our model allows fractional weights, we may not be able to talk about individual items, and thus reneging is not required to be independent between items of the same class. For the simplicity sake's, we assume items arrive at discrete time. We show that stability depends on the exact nature of reneging, in stark contrast with the non-reneging case where it depends on the arrivals only through the arrival rates. To our best knowledge, this is the first such sensitivity result in stochastic matching. We uncover a new stabilising mechanism which exists neither in the non-reneging case nor in the graph case, which explains why incorporating reneging in stochastic matching is not straightforward. Together with balancing mechanism as hinted in the online assignment framework for the non-reneging case, it gives a criterion necessary and sufficient for stability. Finally, whilst verifying stability is a hard problem, we give a family of MaxWeight-type policies parameterised by $\varepsilon > 0$, which are maximally stabilising for all $\varepsilon$ sufficiently small. Unfortunately there is no effective bound for $\varepsilon$, but we show how to adjust its value during implementation.
Read the full abstract
Number Theory
Periodicity and Period-Length Bounds for Browkin $p$-Adic Continued Fractions
Continued fractions have been introduced and studied in the field of $p$–adic numbers by several authors, with the aim of developing analogues of the classical theory of real continued fractions. In this paper, we focus on Browkin's continued fraction algorithm, for which several fundamental questions remain open, especially concerning periodic expansions of quadratic irrationals. In this paper, we solve a conjecture about periodicity left open in [6]. As a consequence, we prove that, for every positive integer $t$, there exist infinitely many square roots of integers whose Browkin continued fraction expansion is periodic with period length $2t$. This also settles a problem originating from previous classical results on the possible period lengths of square roots of integers. Moreover, we provide new sufficient conditions for the periodicity of quadratic irrationals and derive explicit bounds for the corresponding period lengths. These results also restrict the possible behaviour of a quadratic irrational whose Browkin continued fraction expansion is non-periodic, and therefore provide new tools for investigating whether an analogue of Lagrange's theorem can hold. Finally, we also provide a computational study of the Browkin expansions of quadratic irrationals, with particular attention to the behaviour of square roots of integers and to the search for possible non-periodic examples.
Read the full abstract
Artificial Intelligence
Option-Aware Retrieval and Task-Specific VLM Adaptation for Medical VQA
We describe our submission to the MedReason 2026 challenge, covering multiple-choice (MCQ) and open-ended (OE) medical visual question answering (VQA) under fully offline, containerized inference. Our first finding is that MCQ retrieval must compare answer semantics rather than answer labels: labels are independently assigned per question, so copying a retrieved neighbor's label transfers no useful information, whereas scoring each current option's text against correct-answer text from similar training cases raises retrieval-only accuracy from 20.0% to 57.5% on a 200-case retrieval-excluded development holdout. Our second finding attributes the submitted system's accuracy: holding the task-specific MCQ Low-Rank Adaptation (LoRA) adapter fixed and varying the number \(k\) of in-prompt retrieved examples changes accuracy by at most one case — 187/200 (93.5%) at both \(k=0\) and the adapter's training-time \(k=1\), 188/200 (94.0%) at the packaged runtime's default \(k=3\) — and the submitted confidence-gated override adds no net accuracy on top of \(k=3\), selecting the VLM in 198/200 cases. With the final MCQ adapter fixed, retrieval changes accuracy by at most one case, and gating provides no net gain. On 20 OE cases, token-F1 and RaTEScore~\cite{zhao2024ratescore} decrease as \(k\) grows, but paired sign tests on token-F1 differences are nonsignificant (\(p \ge 0.29\)); a single-annotator comparison found 6/20 wrong-anchor errors for the final configuration and 14/20 for an earlier configuration that jointly differed in routing, adapter, and prompting. The system reaches 94.0% MCQ accuracy on the development holdout and 93.20% on the organizer's official pre-evaluation, versus 29.43% for the off-the-shelf reference baseline, while both of the organizer's open-ended scores are lower than that baseline's (ground-truth agreement 1.245 versus 1.588, visual accuracy 1.995 versus 2.696, each out of 4).
Read the full abstract
Instrumentation and Methods for Astrophysics
Gnuastro: removing extended/diffuse outliers while coadding
Deep astronomical images are produced by the coadding (also called "stacking") of several individual exposures. However, individual exposures are often plagued with many artifacts which will reduce the quality of the deep image if not properly removed. Most are removed during reduction/calibration, but the more diffuse may stay. To address this problem, the following new operators were introduced to the Arithmetic program in Gnuastro 0.23: 'madclip-maskfilled', 'sigclip-maskfilled' and 'collapse-*clip-fill-*'. They are able to mask many such outliers successfully by using all the exposures to find outlying pixels in each exposure and exploiting their contiguity. This research note is reproducible with Maneage, on the Git commit bf3815a.