Peer Revenue Publish or profit

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Thursday 3 September

Machine Learning

Local Updates, Global Learning (LUGL): Playing Games with non-incremental Learners

The dominance of Neural Networks (NNs) in RL is partially due to their incremental learning capability, which naturally suits the online, non-stationary nature of self-play training. However, gradient-boosted trees like LightGBM are widely recognised as the state of the art for tabular data in supervised learning, often outperforming NNs in accuracy and efficiency. Game states are inherently tabular—discrete actions, categorical card identities, structured board positions—which makes them an ideal candidate for tree-based methods. We introduce LUGL (Local Updates, Global Learning), a framework that decouples data collection from model fitting, enabling non-incremental learners such as GBTs to operate in RL settings where they would otherwise fail due to distributional shift. LUGL alternates between a local updates phase, where the agent plays self-play games and accumulates tabular updates (Q-values, V-values, policies, or regret values) in a finite table, and a global learning phase, where the table is used to train a function approximator that generalises to unseen states before the table is reset. We test our approach in four standard perfect-information games (Tic-tac-toe, Connect-4, Othello, and Hex) and five imperfect-information games (Kuhn's poker, Leduc Hold'em, Liar's Dice, Goofspiel, and Flop5 Hold'em), and show that our results are competitive with or superior to DQN and DeepCFR. Our experiments demonstrate that the community's strong bias towards NNs in game-playing may be unwarranted, since LightGBM-based agents achieve competitive or superior performance across all tested benchmarks.
Read the full abstract

All authors David Milec, Spyridon Samothrakis, Michael Fairbank, Dennis J.N.J. Soemers

Cryptography and Security

Security and Privacy in the Musical Metaverse: Threat Analysis and Design Implications

The Musical Metaverse (MM) introduces immersive, real-time environments for collaborative musical interaction, characterized by ultra-low-latency constraints, continuous multimodal data streams, and heterogeneous devices. These properties create a distinctive security and privacy landscape that differs significantly from conventional XR or multimedia systems. This paper presents a multi-layer threat analysis of MM ecosystems, identifying key assets including live musical content, expressive interaction data, identity and session metadata, and intellectual property. Threats are analyzed across network, application, data/AI, device, intellectual property rights, and social layers, with particular attention to risks arising from expressive and neurophysiological data, which enable inference, re-identification, and potential privacy violations. We describe a stakeholder-driven survey involving 14 participants from 13 organizations, revealing that neurophysiological data leakage and real-time stream disruption are perceived as the most critical risks, followed by intellectual property infringement and avatar impersonation. We further evaluate the suitability of existing security protocols under strict latency constraints, showing that conventional approaches such as TLS over TCP are often incompatible with real-time musical interaction, while lightweight, stream-oriented mechanisms (e.g., SRTP, DTLS) provide a more suitable balance between security and performance. Based on these findings, we derive a set of design guidelines for MM systems, emphasizing latency-aware security, differentiation of interaction paths, data minimization, and edge-centric processing. The results support a security-by-design approach that enables trust and compliance without compromising real-time performance.
Read the full abstract

Atmospheric and Oceanic Physics

Radiative and Dynamical Controls on the Land-Ocean Warming Contrast in Climate Models

Surface air over land warms substantially more than over the ocean under greenhouse forcing, a phenomenon known as the land-ocean warming contrast. Current explanations for this contrast are commonly expressed either in terms of energetic constraints, from top-of-atmosphere and surface energy balance, or dynamical constraints, from large-scale atmospheric dynamics. We show that these perspectives are complementary when viewed through the lens of atmospheric moist static energy (MSE) transport, and that connecting them yields new insight into the controls of the warming contrast and the spread in climate models. We use this framework to construct an interpretable emulator that reproduces the land-ocean warming response across 22 models from the latest Coupled Model Intercomparison Project (CMIP6). We find that the strength of the land-ocean warming contrast emerges from the interplay between a model-dependent radiative baseline and a robust dynamical restoring mechanism that favors greater warming over land. This interplay leads to two broad model regimes that align with climate sensitivity. In low-climate-sensitivity models, more stabilizing radiative feedbacks over the ocean directly favor greater land warming. In high-climate-sensitivity models, radiative feedbacks alone would instead favor greater ocean warming, but a strong MSE-transport feedback (approximately 0.2 PW/K) more than compensates for this tendency. The intermodel spread in the land-ocean warming contrast is closely related to the ratio of radiative feedbacks over land and ocean, highlighting a broader connection between climate sensitivity and the land-ocean warming contrast.
Read the full abstract

All authors Paolo Giani, Arlene M. Fiore, Raffaele Ferrari, Paul A. O'Gorman, Vincent T. Cooper, Noelle E. Selin

Computer Vision

Rethinking 3D Noise: Learning 3D-Aware Video Priors via Optimization-Free Morphological Perturbations

3D scene representations like NeRF and 3D Gaussian Splatting (3DGS) suffer severe artifacts in sparse-view settings. Recent generative 3D artifact fixers attempt to address this, but rely on paired corrupted and clean renders requiring costly, per-scene reconstructions across varying view configurations. While 2D image augmentations act as instant regularizers, no explicit equivalents exist for 3D representations to preserve spatial consistency across views, an essential property for 3D-aware training. We propose 3D Morphological Perturbations as an optimization-free regularizer that preserves spatial consistency. Leveraging explicit 3DGS, we treat each Gaussian as a fundamental building block - analogous to a 2D pixel - and apply perturbations across its morphological parameter space via scale, rotation, and pruning. Our method eliminates per-scene 3DGS optimization loops from dataset curation while enabling models to learn stronger geometric priors than sparse-view baselines in diagnostic ablations conducted on a lightweight video diffusion sandbox. Scaled to a 14B-parameter video model via ControlNet, our approach maintains visual fidelity while reducing mean depth error by 12.5% over state-of-the-art image-to-image 3D artifact refiners, ultimately boosting downstream robotics policy success rates by up to 8.0% across 3 of 4 manipulation tasks.
Read the full abstract

All authors Onat Şahin, Mohammad Altillawi, George Eskandar, Carlos Carbone, Ziyuan Liu

Quantum Gases

Finite-temperature mass gap and quench dynamics of mobile impurities in a Fermi gas

Recently, a mass-gap description of mobile impurities in a Fermi gas was introduced, which connects Anderson's orthogonality catastrophe for static impurities to the quasiparticle picture of Fermi polarons through a recoil-induced energy gap in the fermionic dispersion. That description, however, was restricted to zero temperature and did not address dynamics. Here we generalize the mass-gap model to finite temperature by combining the Lee–Low–Pines transformation with a self-consistent Hartree–Fock decoupling of the recoil-induced interaction, and we study the quench dynamics within this framework using the functional-determinant approach. At finite temperature the effective mass gap obeys the self-consistency equation $\Delta(T)=2U_F\tanh[\Delta(T)/4k_B T]$, with $U_F=k_F^2/2M$ and $M$ the impurity mass. This equation admits a nonzero solution below the characteristic temperature $T^*=U_F/(2k_B)$ and closes as $(T^*-T)^{1/2}$. We identify this closing as the mean-field signature of the thermal melting of the polaron and molecule quasiparticles. Computing the Ramsey response $S(t)$ after a sudden quench of the impurity–fermion interaction, we find that its long-time oscillations—quantum beats between the bound and in-gap states—disappear precisely above $T^*$. Our work ties the thermodynamic and dynamical fingerprints of polaron formation to a single temperature-dependent mean-field parameter.
Read the full abstract

Computer Vision

PL-SCEA: Reconfiguring Pretrained Attention for Few-Shot Industrial Anomaly Detection

Vision Foundation Models (VFMs) provide transferable patch representations for few-shot industrial anomaly detection, but their attention computation is typically inherited from pretraining objectives centered on semantic aggregation. This creates a potential mismatch: token relations that support semantic recognition may not adequately expose the localized texture and structural deviations required for anomaly localization. We therefore investigate the hypothesis that the attention computation of a frozen VFM can be reconfigured as a task-relevant component of anomaly detection. We instantiate this idea with Power-Law Self-Correlation Enhanced Attention (PL-SCEA), which retains the semantic context of pretrained query-key attention while constructing token-adaptive self-correlations over contextualized value features. Positive-correlation filtering and power-law reweighting then emphasize relations that are salient relative to each token's relational background, without introducing additional trainable attention projections. The resulting features are modeled by a lightweight variational autoencoder that provides a fixed-size reconstruction-based representation of category-specific normality. The two stages serve complementary roles: attention reconfiguration shapes how local relational deviations are represented, while reconstruction-based modeling converts deviations from learned normality into anomaly scores. Across MVTec AD and VisA, the complete framework achieves competitive image-level detection and consistently strong pixel-level localization across the evaluated few-shot settings. Ablations further show that PL-SCEA improves localization with either the VAE or a memory bank under the tested setting. These results support the view that task-aligned attention reconfiguration can improve the anomaly-localization capability of frozen pretrained representations.
Read the full abstract

All authors Xiaoyu Yang, Qixing Wu, Huixian Zhao, Changlong Jin

Computation and Language

Enhancing Financial Question Answering: A Novel Benchmark Dataset of Banks' financial statements

The comparative analysis of banks' financial statements poses significant challenges for automated question answering systems due to their complexity, substantial length, technical language, and inhomogeneity of both textual and numerical content across different jurisdictions and institutions. We introduce FinRAG-QA, a novel benchmark dataset for financial question answering, which comprises 999 practitioner-curated questions on 10 standardised indicators, grounded in 209 annual and Pillar 3 reports from 24 major European and U.S. banks spanning 2019-2023. Unlike prior financial QA benchmarks, which centre on U.S. filings and single-institution analysis, FinRAG-QA targets cross-institutional retrieval over documents averaging 198k words, longer than any existing financial QA resource. On this benchmark we evaluate a multi-stage RAG pipeline and isolate the contribution of each component. Contextual chunk enrichment combined with a retrieval-optimised embedding model raises NDCG@10 from 0.322 to 0.710; conditional on the ground truth being retrieved, a reasoning-optimised generator raises answer accuracy from 44.6% to 79.0% (+34.4 percentage points), at roughly 20x the generation latency. We further show that cross-encoder reranking degrades retrieval when the first-stage ranking is already strong, and that a single top-ranked chunk outperforms larger contexts at generation time. Experiments were run in late 2024-early 2025 with the models available at that time.
Read the full abstract

All authors Arianna Miola, Bruno Spaccavento, Lorenzo Silotto, Marco Bianchetti, Luca Cagliero

Number Theory

The derived set of multiple star polylogarithms

The derived set of multiple zeta-star values is the half-line $[1,+\infty)$. In this paper we study the corresponding two-dimensional problem for multiple star polylogarithms on the unit circle. We first prove that every shifted multiple star polylogarithm is the generating function of a completely monotone sequence and hence admits a normalized Hausdorff–Stieltjes representation. Both the shifted and the unshifted functions, as well as their infinite-depth limits, are shown to be univalent on the half-plane $\mathrm{Re}\,z<1$. For finite indices, the representing densities satisfy a strict monotone likelihood-ratio order with respect to the reverse lexicographic order. Together with closure properties of the Hausdorff–Stieltjes class and a limiting quotient argument, this shows that, along the upper semicircle, the argument of each translated curve increases strictly while its modulus decreases strictly; the corresponding lower-semicircle statement follows by conjugation. The same idea gives strict separation of the curves attached to different indices. We then establish a one-to-one correspondence between a natural set of pairs consisting of a point of the unit circle and an infinite index, and the closed half-plane $\mathrm{Re}\,w\geq \frac12$ with the point $1$ removed. Under a binary coordinate on the index set, this correspondence is a homeomorphism. As an application, a subclass of shifted cyclotomic multiple zeta-star values of all levels form a countable dense subset of the open half-plane $\mathrm{Re}\,w>\frac12$. Consequently, the derived set, and every higher derived set, of this cyclotomic family is the closed half-plane $\mathrm{Re}\,w\geq\frac12$.
Read the full abstract

Computation and Language

The Impact of Synthetic Data Augmentation on Discourse-Pragmatic Function Classification

Synthetic data augmentation has become a common strategy for addressing class imbalance in NLP, but most approaches focus on the quantity and diversity of generated examples rather than their geometric relationship to real training data. We investigate this question in the context of discourse pragmatic function classification, a task where data sparsity is a structural feature rather than a collection artefact. Using 410 manually annotated instances of the English word look drawn from the British National Corpus, spanning four functions: Attention Signal, Directive, Discourse Marker, and Interjection. We generate synthetic training examples with Llama 3.1 and partition them by their cosine distance from real training data in RoBERTa embedding space. We compare six training conditions that differ in the placement of synthetic examples relative to the empirical decision boundary, while holding augmentation quantity constant across conditions. All augmented conditions improve macro F and accuracy over the real only baseline, but core proximal examples (NEAR) yield the largest gains in macro F (0.113), while a distance balanced mix achieves the highest accuracy (0.748). No condition improves AUC, indicating that augmentation shifts the decision boundary rather than improving the model's underlying probability estimates. These findings suggest that where synthetic examples land in representation space matters as much as how many are generated, with implications for low resource pragmatic classification more broadly.
Read the full abstract

Representation Theory

Jantzen filtrations for graded affine Hecke algebras

We prove the Jantzen conjecture for standard modules of graded affine Hecke algebras with equal parameters, real central character in the dominant deformation direction. The Jantzen layers are shown to be semisimple, and the multiplicities are given by the full local intersection cohomology of the corresponding orbit closures. The proof identifies the algebraic standard-to-contragredient map with the costalk-to-stalk map on a contracting transverse slice. Hard Lefschetz theorem determines its elementary divisors, and a residual grading on the specialized convolution algebra proves semisimplicity. An example in type $A$ is included to show that dominance is necessary.
Read the full abstract

Mesoscale and Nanoscale Physics

Low-Frequency Charge Noise in Bilayer Graphene Quantum Dots

Bilayer graphene (BLG) quantum dots (QDs) are a promising platform for semiconductor qubits. However, the low-frequency charge noise that may ultimately limit coherence has remained largely unexplored. Here, we systematically characterize charge noise in gate-defined BLG QDs using transport-based noise spectroscopy. We extract a median amplitude of $ S_\mu^{1/2} (1~\text{Hz}) = 1.16~\mu\text{eV}/\sqrt{\text{Hz}}$, placing BLG well within the range reported for established semiconductor quantum-dot platforms. Across variations in charge occupation, confinement, source-drain bias, and charge-sensor operating conditions, neither the noise amplitude nor the spectral dependence shows a reproducible trend in electrostatic tuning, indicating that we extracted the intrinsic semiconductor noise. Consistent noise levels are further observed in double QDs and confirmed using an independent superconducting resonator-based dispersive readout. Extending the study to BLG devices incorporating transition metal dichalcogenide layers reveals no measurable charge noise increase in weakly proximitized QDs. These results validate BLG as a viable platform for coherent quantum information processing.
Read the full abstract

All authors Jessica Richter, Max J. Ruckriegel, Jonas D. Gerber, Tijl Degroote, Christoph Adam, Markus Niese, Lara Ostertag, Clara Galante, Kenji Watanabe, Takashi Taniguchi, Petar Tomic, Artem O. Denisov, Hadrien Duprez, Klaus Ensslin, Thomas Ihn

Cosmology and Nongalactic Astrophysics

Prospects for probing dark matter with filamentary 21cm emission

The cosmic web contains most of the matter in the Universe, and its filamentary structure is closely related to the properties of dark matter. Compared with the cold dark matter (CDM) model, warm dark matter (WDM) suppresses the formation of low-mass halos and produces smoother cosmic filaments. In this work, we explore whether filamentary 21cm emission can be used to distinguish between these two dark matter models at $z=4$ and $2.5$. We generate intrinsic 21cm brightness temperature maps from cosmological hydrodynamical simulations and produce mock observations by convolving them with Gaussian beams of different angular resolutions. We then compare the predicted filament signals with the expected sensitivities of SKA1-Low and SKA2-Low. We find that the intrinsic 21cm emission clearly reflects the different filament morphologies in the CDM and WDM models, particularly at $z=4$. However, these differences rapidly disappear as the beam size increases, especially for beam sizes larger than 6 arcsec. Although larger beams improve the observational sensitivity, they also erase the small-scale structures that encode the dark matter information. As a result, neither SKA1-Low nor the planned SKA2-Low can directly resolve individual filaments with sufficient sensitivity to distinguish between the two models. We therefore conclude that direct imaging of cosmic filaments and constraining dark matter through their morphology are unlikely to be feasible with current and planned SKA-Low facilities.
Read the full abstract

All authors Yizhou Liu, Yingjie Jing, Haoran Shi, Huijie Hu

High Energy Physics (Phenomenology)

Light Leptoquarks in a Dark Sector: Scalar Dark Matter, Neutrino Mass, and Collider Signatures

We investigate the dark matter (DM) phenomenology and subsequent collider signatures of a dark leptoquark (LQ) model containing dark vector-like quarks (VLQs) and a scalar singlet. The dark LQs and VLQs participate in the radiative generation of Majorana neutrino masses at one loop. Since the coloured LQs and VLQs cannot serve as viable DM candidates, a $\mathbb{Z}_2$-odd singlet scalar is introduced as the DM candidate. The $\mathbb{Z}_2$-odd nature of the LQs forbids their conventional decays into a quark and a lepton, allowing them to evade the standard LHC constraints that apply to visible LQ signatures. This framework therefore offers the distinctive possibility of sub-TeV dark LQs coexisting with TeV-scale dark VLQs. These sub-TeV dark LQs help in achieving the observed relic abundance of the singlet through co-annihilation. We analyse the resulting DM phenomenology and assess the prospects for probing this scenario at a future muon collider.
Read the full abstract

All authors Priyotosh Bandyopadhyay, Kirtiman Ghosh, Anirban Karan, Snehashis Parashar, Debabrata Sahoo

Analysis of PDEs

Extension Problems for the Vladimirov–Taibleson Operator and the Hierarchical Laplacian

We establish a non-Archimedean Caffarelli-Silvestre extension theory for the Vladimirov-Taibleson operator $D^s$ for $s>0$ and, more generally, for hierarchical Laplacians $L_C$. For every $f\in\mathcal{S}(\mathbb{Q}_p^n)$, we equip the Bruhat-tits tree $\mathcal{T}_{p^n}$ with the weight $w_{v_{k-1}v_k}=p^{k(s-n)}$ under the horocyclic coordinates. We prove that the resulting Dirichlet problem admits a unique bounded weighted-harmonic solution that is continuous on the end compactification. Its boundary traces converge to $f$ uniformly and in $L^2$, while the associated deformed normal derivative converges pointwise and in $L^2$ to $D^s f$. We derive explicit Fourier and Poisson representations of the extension and establish an energy identity between its weighted tree energy and the quadratic form of $D^s$. Motivated by $p$-adic AdS/CFT, we also give an equivalent massive formulation on the unweighted tree, based on a renormalized trace and a scale-corrected normal derivative. Finally, under natural local-finiteness and scaling assumptions, we extend the construction to hierarchical Laplacians $L_C$ on ultrametric spaces. We characterize the canonical edge conductances by requiring the cancellation Green operator on the ultrametric tree $\mathcal{T}_X$ to coincide with $L_C^{-1}$ on $\mathcal{S}_0(X)$. In the corresponding canonical flux class, the extension is unique, its Dirichlet-to-Neumann map is $L_C$, and it satisfies the associated energy identity. The Vladimirov-Taibleson construction is recovered as the homogeneous special case.
Read the full abstract

High Energy Physics (Phenomenology)

A Bayesian Pixel Based Approach for Model Independent TMD Reconstruction

We introduce a nonparametric pixel-based framework for the Bayesian inference and imaging of transverse momentum dependent (TMD) parton distributions. The methodology integrates TMD evolution within the Collins-Soper-Sterman formalism in a differentiable framework, and leverages generative AI through a hybrid normalizing flow-driven Metropolis-Hastings algorithm for efficient posterior sampling. The framework is validated through multi-scale closure tests of increasing complexity. Using singular value decomposition, we characterize the existence of null TMDs, functional components that remain unconstrained by observables, and demonstrate how multi-scale data break these degeneracies, enabling 3D partonic imaging.
Read the full abstract

Computational Engineering

Calibration of neural viscoelastic models via full-field data

We propose an unsupervised learning framework for calibrating a physics-augmented neural network (PANN) for small-strain viscoelasticity via full-field data. It only requires quantities that are directly accessible in real experiments for training, namely global reaction forces and surface displacements. The underlying PANN is embedded in the generalized standard materials theory, in which two scalar-valued potentials render the constitutive model thermodynamically consistent by construction, while invariant-based representations of the free energy and the dual dissipation potential additionally ensure material symmetry. Considering a thin specimen under the plane stress assumption, we formulate a constrained optimization problem based on the equilibrium gap method in combination with quasi-Newton optimizers and automatic differentiation. Thereby, the unknown out-of-plane strain follows from the plane stress condition and the evolution of the internal variables is captured by an implicit time integration scheme. The resulting system of nonlinear equations is solved via a local Newton iteration at quadrature point and time step. To drastically reduce the computational cost of training, the backward adjoint method is employed to compute the gradient of the target loss, instead of backpropagating through all Newton iteration steps. The proposed framework is demonstrated for synthetic data, including noisy displacements and forces, showing excellent agreement across a wide range of deformation rates and load paths.
Read the full abstract

Medical Physics

HPC Modeling of Coupled Elastic-Acoustic Wave Propagation in Biological Media: Numerical Validation

Accurate numerical models of sound propagation through biological media are an important tool for many applications, from medical physics to studying the auditory system of humans or other animals. We model high-frequency elastic and acoustic wave propagation through the head anatomy of a common bottlenose dolphin (Tursiops truncatus), by means of the open-source software packages specfem3d, based on the spectral-element method, and \textsc{k-wave}, based on the pseudospectral method. To achieve sufficiently high performance, we ported the latter solver to C++, for Multi-GPU CUDA support via the slab-decomposition of three-dimensional Fast Fourier Transform approach. Because the two schemes differ fundamentally in how the propagation medium is discretized and internal (in particular fluid-solid) interfaces are treated, similarity between modeled signals across methods is a legitimate measure of model accuracy. Plane waves, depending on time like four-cycle sinusoidal bursts of varying central frequency (20–100 kHz), are numerically propagated through a computed-tomography-based anatomy model using both solvers. The sound is “recorded” in front of the rostrum and at the right inner-ear locations for comparison. We successfully cross-validate both solvers on high-fidelity High-Performance Computing (HPC) clusters. We find that the stability of results from both solvers grows as the corresponding spatial resolution is refined. At their highest resolutions, the two methods show excellent agreement, with normalized correlation exceeding 0.99 across the entire 20–100 kHz frequency range. Together, our results provide a validated HPC-simulation framework for wave propagation in biological media, with broader implications, e.g., for biosonar research, auditory biomechanics, and medical ultrasound.
Read the full abstract

Formal Languages and Automata

Relative Prime Factorization and Finite-State Presentations under Fixed Finite-Monoid Observation

Let $L\subseteq\Sigma^*$ and fix a morphism $h:\Sigma^*\to M$ into a finite monoid. We study exact factorization and canonical presentation in the relative syntactic congruence $\theta_{L,h}:=\equiv_L\cap\ker h$. We separate unique factorization from finite direct presentation. An exhaustively computer-checked $36$-element quotient has a unique exact prime factorization for every live non-unit class, yet its valid prime-return rules contain an infinite family, so unique factorization does not imply the finite relative presentation property (FRP), even for a finite quotient. We lift the same defect to a nonregular context-free language with an infinite relative quotient and finite prime spectrum. To isolate the obstruction, we introduce the finite-state relative presentation property (FSRP), in which canonical valid right-hand-side languages are represented by finite residual controllers, and prove $\mathrm{FRP}\subsetneq\mathrm{FSRP}$. We then introduce prime-target left-division determinism (PTLD), which implies unique exact factorization, tail exactness, tail determinism, and a quadratic bound on valid rules. A nonregular deterministic context-free example with a finite group observer satisfies PTLD while lying outside every fixed $(k,\ell)$-substitutable class. Finally, for fixed $h$ we give a strong positive-data learner for the canonical PTLD presentation with polynomial-time hypothesis updates and a finite characteristic sample, together with a limit reconstruction of the canonical FSRP controller from weakly behaviorally correct CFG-valued learners.
Read the full abstract

Optics

Machine learning-assisted design and explainable optimization of CdSnP2-based integrated solar-photodetector devices

CdSnP2-based integrated solar cell-photodetector (SC-PD) devices employing CdS and CuGaSe2 (CGS) as the window and back surface field (BSF) layers, respectively are investigated using a hybrid machine learning (ML)-assisted SCAPS-1D framework. Device optimization is performed by varying the thickness, doping concentration, and defect density of individual layers. SCAPS-generated data are used to train six ML models and one deep learning model, with ensemble-based algorithms exhibiting the highest predictive accuracy. The ML-guided optimization identifies the n-CdS/p-CdSnP2 (CTP)/p+-CGS architecture as the optimum configuration among thirteen candidate structures. Incorporation of a 200 nm CGS BSF layer significantly enhances both photovoltaic and photodetection performance, increasing the efficiency from 20.67% to 32.69%, responsivity from 0.53 AW-1 to 0.72 AW-1, and detectivity from 2.51x1014 Jones to 1.78x1016 Jones. SHapley Additive exPlanations (SHAP) analysis reveales that band-offset engineering, particularly at the window/absorber and absorber/BSF interfaces, together with absorber properties, governs device performance. These findings demonstrate the potential of CdSnP2 and the proposed data-driven SCAPS-ML framework for the accelerated design of high-efficiency multifunctional optoelectronic devices.
Read the full abstract

All authors Md. Alamin Hossain Pappu, Kazi Abrar Shafin, Mainul Hossain, Jaker Hossain

Computer Vision

Tree-VQ: Progressive Image Compression from Pretrained Vector Quantizers

Progressive image compression requires a single embedded representation whose received prefixes can be decoded without re-encoding the source. Modern vector-quantized (VQ) image models provide strong discrete endpoint representations, but conventional flat codeword indices do not define meaningful intermediate states for a neural decoder. We present Tree-VQ, a post-hoc conversion of a pretrained flat VQ tokenizer into a fine-grained, arbitrary-prefix progressive representation while preserving its encoder assignments and every learned leaf vector. The key idea is to organize the original codebook into a balanced binary hierarchy, associate explicit representations with internal nodes, and transmit branch decisions in depth-major order. Consequently, once the image header is available, every payload prefix uniquely specifies a valid latent state: each additional branch bit refines exactly one token, and transmission can therefore be truncated at essentially any payload position rather than only at a small number of stage boundaries. We further adapt one shared decoder on the complete-depth and mixed-depth latent states encountered under such arbitrary truncation, making these densely spaced prefixes useful for reconstruction rather than merely syntactically decodable. On Kodak, Tree-VQ achieves a DISTS-based BD-rate saving of 52.1% relative to ProGIC, while exposing thousands of valid arbitrary-prefix operating points from a single embedded bitstream.
Read the full abstract

All authors Mingming Ma, Xinkun Wang, Tianyi Xu, Qingyu Luo, Fu Li, Yi Niu

Quantum Physics

Quantum Hamiltonian Evolution for Coherent Quantum Learning

We introduce Coherent Quantum Learning (CQL), a training framework for quantum learning models in which the model parameters are quantum degrees of freedom evolved under a Hamiltonian that encodes the loss function. Current quantum machine learning retains classical optimization: parameters are updated by a classical outer loop using gradient estimates from measurements, and quantum coherence has no role in the training dynamics, just as in any classical treatment of the same problem. In the quantum case, a parameter register initialized in superposition evolves unitarily, and probability amplitude concentrates near low-loss configurations through interference, without gradient computation or classical feedback. We give an explicit construction using block encodings and Hamiltonian simulation, applicable to arbitrary parameterized circuits. Numerical experiments on binary classification and interferometric phase estimation confirm that the evolved distribution peaks at the optimal parameters, matching gradient-based performance. The construction is compatible in principle with fault-tolerant implementations and extends to batched training via sequential Hamiltonian evolution.
Read the full abstract

All authors Ignacio B. Acedo, Javier Gonzalez-Conde, Pablo Rodriguez-Grasa, Barry C. Sanders, Lirandë Pira

Computer Vision

Stabilizing Camera-Controlled Novel View Synthesis at Inference Time

Training-free, camera-controlled novel view synthesis from a single image using pre-trained video diffusion models often becomes unstable under large camera motion and long generation horizons. Existing approaches commonly combine several inference-time components, making it unclear which design choices are most important for stability. This work identifies the amount of camera motion handled by each diffusion pass as the main factor. CamTrol++ splits the camera path into passes that each cover at most a fixed rotation and generates them one after another, so that every pass stays within the range of motion the backbone handles well. A matched study, in which every setting is scored on the same poses of the same $60^\circ$ orbit, shows that geometric error grows steadily with the arc per pass, while a single SVD pass both loses geometry and tends to freeze the camera. Under one shared camera path, CamTrol++ achieves the lowest epipolar error and the best held-out 3D Gaussian Splatting reconstruction among CamTrol, NVS-Solver and WAVE, and reaches the quality of CamTrol with the same pass structure at least $3.2\times$ faster. On real RealEstate10K motion it matches CamTrol at half the run time (NVS-Solver is more faithful but $8\times$ slower). The same procedure also improves 3D reconstruction with a second video backbone, LTX-Video. Geometry-constrained spatial attention, color anchoring at pass boundaries and a registration-free warping pipeline that makes warping $6.1\times$ faster complete the framework. The method remains effective for 56-frame generation and under controlled depth corruption. These results show that careful control of camera motion at inference time can substantially improve the stability of camera-controlled novel view synthesis.
Read the full abstract

All authors Prajwal Singh, Arjun Badola, Seema Kumari, Hajime Nagahara, Shanmuganathan Raman

Quantum Physics

Entanglement dynamics of accelerated atoms with environment-induced interactions

We investigate the influence of environment-induced interactions on the entanglement dynamics of two uniformly accelerated atoms coupled to a fluctuating massless scalar field with a reflecting boundary. The two atoms are aligned perpendicular to the boundary. The entanglement behavior is influenced by the competition between the environment-induced interatomic and the environment-induced atom-plate interactions, which can be characterized by certain critical values. The maximum of concurrence generated during evolution decreases non-monotonically with the acceleration, which implies that the anti-Unruh phenomenon can exist for some situations even when both environment-induced interatomic and environment-induced atom-plate interactions are considered.
Read the full abstract

Functional Analysis

On the abstract approach to spectral constants: a proof of the Clouâtre–Ostermann–Ransford conjecture

Clouâtre, Ostermann, and Ransford formulated an abstract version of Crouzeix's conjecture involving a bounded unital homomorphism from a uniform algebra into matrices and a unital antilinear map. They conjectured that contractivity of the associated symmetrised map forces the homomorphism to have norm at most two. We prove this conjecture, in fact without requiring the antilinear map to be contractive and for homomorphisms into the bounded operators on a Hilbert space. The proof combines positivity of real parts, an operator-valued Herglotz theorem, and the perturbation lemma of Lorist and Schwenninger used in the recent proof of Crouzeix's conjecture.
Read the full abstract

Nuclear Experiment

Measurement of the forward angle $^{12}$C+$^{12}$C fragmentation differential cross sections at 62 MeV/nucleon

The present work reports on high-precision measurements of forward-angle fragmentation differential cross sections for the [12]C +[12] C reaction at 62 MeV/nucleon using the FAZIA array. Angular distributions for fragments from Z = 1 to 6 were extracted in the range 2 <= theta_lab <= 8 degrees. Particle identification was achieved by combining the Delta E - E technique with Pulse Shape Analysis, and precise energy calibrations were performed. The results show that for heavier fragments, the angular distributions are better described by the Van Bibber formulation than by the Goldhaber model, consistent with the dominance of a wide component from dissipative processes in the measured angular range. Notably, alpha particles exhibit an anomalously narrow angular distribution, likely originating from the intrinsic cluster structure of [12]C. Comparisons with existing data at the same incident energy show very good agreement, with a more complete set of species reported here, while a kinematic scaling is proposed to compare our data with previous experimental data at 50 and 95 MeV/nucleon.
Read the full abstract

All authors G. Guo, G. Casini, B. H. Sun, S. Barlini, A. Camaiani, C. Frosin, I. Lombardo, O. Lopez, S. Piantelli, I. Tanihata, S. Terashima, S. Valdré, G. Verde, F. W. Zhao, L. Baldesi, B. Borderie, R. Bougault, C. Ciampi, I. Dekhissi, J. A. Dueñas, Q. Fable, F. Gramegna, D. Gruyer, A. Hocine, B. Hong, S. Kim, T. Kozik, S. H. Nam, N. Le Neindre, J. Park, M. Parlog, E. Vient, M. Vigilante, C. J. Wang

Artificial Intelligence

Analysis of Prompt Engineering for Drug Toxicity Prediction

Clinical trials in the UK can cost up to {\pounds}1.3 million, with approximately 90% drug failure rate. Toxicity is a major contributing factor in drug failure. Testing is time and cost intensive. In recent years, the use of artificial intelligence has been increasingly explored to aid in the prediction of drug toxicity, with extensive use of large language models (LLMs). However, LLMs can show considerable variation when minor changes are made to prompts, which raises concerns about their sensitivity to prompt engineering. Prompt engineering is used to optimise a prompt given to an LLM to generate the desired output. This paper proposes a method to analyse prompt engineering for drug toxicity prediction. The aim of the paper is to investigate the importance of prompt phrasing for drug toxicity prediction. LLMs were prompted to identify chemical properties of significance when predicting drug toxicity. Prompts were constructed to investigate; job role, prompt structuring, and rule interpretation. LLMs were then used to generate datasets, using the identified features from initial prompting, which were then passed to machine learning algorithms. The experiments show that the natural variance which occurs in LLMs outweighs any fine-tuning of prompts. There were, however, substantial improvements in model performance when using chemoinformatic code to extract features instead of using LLM-generated values. The proposed analysis methodology is applicable to a wide range of prompt types across different areas of bioinformatics.
Read the full abstract

Probability

Joint Laws of Maximum Drawdown and Maximum Drawup for Spectrally Negative Lévy Processes

Let X be a spectrally negative Lévy process observed up to an inde- pendent exponential time T with parameter {\gamma} > 0. We study the joint law of the maximum drawdown and the maximum drawup of X on [0, T ]. The path is decomposed according to the two possible orderings of its in- fimum and supremum. Conditional on the values of these extrema and on their ordering, the resulting pre-, intermediate, and post- components are independent. We identify their laws as Doob h-transforms of killed spec- trally negative Lévy processes and express the corresponding distribution functions explicitly in terms of the {\gamma}-scale functions W ({\gamma}) and Z({\gamma}) and their derivatives. Combining the conditional laws with the joint densities of the extrema yields integral representations of the joint distribution and moments.
Read the full abstract

Computation and Language

</think> Doesn't Stop Reasoning: Analysis of Spurious CoT Termination

Chain-of-thought (CoT) reasoning improves large reasoning models (LRMs) on complex tasks but often produces long, redundant traces. Recent training-free early-exit methods shorten these traces by choosing an intermediate point to stop reasoning. We study one such strategy that injects an end-of-think token (EoT, </think>) at this point to trigger the reasoning-to-answering transition, and find that the injected EoT does not always induce a clean answering phase. Answering-phase generation can continue before the model regenerates another EoT, with the span preceding this regenerated EoT scaling with the reasoning tokens saved by early exit and exhibiting continued reasoning behavior. We call this spurious CoT termination, where reasoning-like generation continues into the answering phase. We hypothesize that insufficient attention to the injected EoT contributes to spurious CoT termination and probe this hypothesis with Exit-token Attention Biasing (EAB). Across four LRMs, five benchmarks, and two early-exit methods, increasing attention to the injected EoT reduces spurious CoT termination and answering-phase length. These results reveal a limitation of controlling LRMs by externally matching their explicit think-block format. Inserting the EoT token conforms to this format but does not by itself guarantee the intended reasoning-to-answering transition. Our code is available at https://github.com/Seunghee-Koh/Spurious-CoT-Termination.
Read the full abstract

All authors Seunghee Koh, Sungjae Choi, Minchan Kwon, Sunghyun Baek, Junmo Kim

Physics and Society

Dynamic probabilistic decision networks

A new type of decision networks is suggested and its operation is analyzed. The network nodes are represented by intelligent agents who can denote either some biological beings, like humans, or neurons of the brain, or the nodes of artificial intelligence. The specifics of the network are in the following: It is probabilistic in the sense that the choice, accomplished by each agent, is characterized by the related probability. It is dynamic, with the probabilities varying in time due to the exchange of information between the agents. It is affective, because the agents choose between alternatives by taking account of utility as well as of biases and emotions. In general, it is heterogeneous, being composed of the groups of agents with different properties, for instance having long-term memory and short-term memory. The network dynamics, caused by the information exchange, results in decision error decrease. The network operation is illustrated by the example starting with the Allais paradox, its resolution, and the decision error diminution in the process of decision dynamics with information exchange. Resorting to machine-learning techniques it is possible to regulate the behavior of the network agents forcing them to choose particular alternatives.
Read the full abstract

Numerical Analysis

A fully globalized solver for discretized inverse elliptic coefficient problems with exact data

We consider finite-dimensional nonlinear inverse problems arising from finite element discretizations of elliptic inverse coefficient problems such as the Calderón problem with finitely many measurements and unknowns. Such inverse coefficient problems are notorious for their nonlinearity and ill-posedness, and numerical solvers tend to depend strongly on good initial values. In this work, we develop a new locally convergent algorithm with an explicit residual criterion that ensures convergence to the inverse problem solution, and a globalized variant that is guaranteed to automatically switch to the faster locally convergent algorithm after finitely many global search steps.
Read the full abstract