Audio and Speech Processing
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Thursday 27 August
We present FDverb, a frequency-domain artificial reverberator, based on the idea of a vocoder with a noise carrier signal. Using a short-time Fourier transform (STFT) for analysis and synthesis, FDverb generates late reverberation by weighting spectral noise components with envelopes. We extend FDverb with early reflections, nonlinear decay, and pitch shifting. These extensions enable creative sound-design applications. We provide FDverb as an open-source DAW plug-in, using the JUCE framework.
Optimization and Control
Solving Quasi-Variational Inequalities Using the Progressive Decoupling of Linkages
Inspired by the progressive decoupling of linkages methodology for optimization and variational inequalities, we propose an algorithm for solving quasi-variational inequalities as a sequence of variational inequalities. Our method is shown to converge locally under some regularity conditions and globally when such conditions hold throughout the entire domain. Separately, under other type of assumptions, global convergence with linear rate is also established for the class of quasi-variational inequalities said to have a moving set. Advantageous computational performance is shown for large-scale Walrasian equilibrium problems, a special case of generalized Nash equilibria, as well as for some other instances encountered in related literature.
Read the full abstract
Signal Processing
Secure Pseudonymetry with DSSS Watermarking for LEO Satellites
Low-Earth-orbit (LEO) constellations offer dense signals of opportunity (SOP) for positioning, navigation, and timing (PNT) without dedicated navigation infrastructure. However, using these signals passively, by only listening to the downlink, is difficult when beacon/pilot components are weak, signals from multiple satellites overlap, or the receiver cannot reliably identify which satellite generated a given observable. This paper proposes a low-power direct-sequence spread-spectrum (DSSS) watermarking framework that embeds a recoverable pseudonymous satellite identifier and keyed authentication field into a LEO downlink signal. The watermark carries a public satellite pseudonym for identification and a keyed HMAC-SHA256-based authentication field for authorized verification, enabling passive detection and verification without full downlink demodulation. We evaluate the framework in a multi-satellite simulation using Starlink two-line-element (TLE)-derived geometry, Doppler, and link-budget scaling with Sionna RT-based local propagation. The results demonstrate candidate-satellite detection, public-ID recovery, and keyed verification under co-observed satellite interference, with limited perturbation to the simulated beacon/pilot and primary downlink components. The framework targets future transmitter-enabled LEO systems while remaining transparent to legacy receivers that do not process the low-power watermark.
Read the full abstract
Atmospheric and Oceanic Physics
Application of the latent twins approach for clear sky retrieval from IASI observations
In recent years, data-driven approaches emerged as alternatives to traditional physics-based retrievals, taking advantage of machine learning techniques such as learnable pseudoinverse, random forests, or deep learning architectures. Classical data-driven models generalize poorly to out-of-sample regimes, as they optimize over finite datasets without incorporating underlying physical laws. This often requires large models and extensive data to achieve reliability. Physics-Informed Neural Networks address this by embedding physical constraints into the learning process, enabling improved extrapolation. However, they requires substantial computational cost due to the need to solve governing equations at each training step. In this work, we introduce a novel deep learning architecture, based on latent twin approach, that balances model complexity, dataset size, and training cost, while providing a quantitative measure of data quality. This architecture is applied to IASI spectra, with the goal to assess the robustness of this method for retrieving atmospheric profiles, including temperature, water vapor, ozone, surface emissivity, and surface temperature, in real-world clear-sky conditions. The algorithm is first applied on synthetic radiances derived from the NWP SAF database using the fast radiative transfer code sigma-IASI/F2N. After validating the architecture on synthetic data, the algorithm is applied to IASI Level 1C observations, along with their corresponding Level 2 products which serve as reference to evaluate the reconstruction accuracy of the autoencoder-based retrieval. The retrieval performances are discussed along with possible strategies to provide an error analysis for the reconstructed thermodynamical profiles.
Read the full abstract
High Energy Astrophysical Phenomena
Spin-disc misalignment drives periodic accretion and pulse profile asymmetry in X-ray pulsars
We investigate magnetospheric accretion in disc-fed X-ray pulsars assuming that the neutron star spin axis is not perpendicular to the disc plane. We focus on X-ray pulsars, where a geometrically thin disc is truncated far from the stellar surface the channelled part of the accretion flow is expected to be guided by the large-scale dipolar magnetic field after coupling to it near the disc-magnetosphere boundary. Using numerical simulations of plasma motion from the inner disc edge to the neutron star surface, we show that a finite inclination between the disc normal and the stellar spin axis leads to periodic modulation of the mass accretion rate onto the magnetic poles even for a steady mass supply through the disc. This purely geometrical effect arises because stellar rotation changes the orientation of the magnetosphere relative to the disc, producing phase-dependent mass loading of magnetic field lines. The amplitude and shape of the modulation are determined by the system geometry and by the ratio of the stellar spin period to the flow time through the magnetosphere. The resulting variability affects the structure and luminosity of emitting regions near the neutron star surface and leads to asymmetric X-ray pulse profiles. Even without intrinsic asymmetries of the emission regions, this mechanism breaks the time-reversal symmetry expected for stationary accretion and naturally contributes to the observed asymmetry of pulse profiles and phase-resolved spectral features. The effect may also be relevant for ULX pulsars, where intrinsic accretion rate modulation can help preserve strong pulsations in the presence of geometric beaming.
Read the full abstract
Image and Video Processing
CARDINAL Predicts Cardiovascular Risk From Non-contrast Cardiac CT
Cardiovascular risk prediction remains limited by incomplete clinical data and imaging biomarkers that reduce computed tomography (CT) to a small number of handcrafted features. We developed CARDINAL (Cardiovascular Assessment via Representation learning from Deep Imaging with Nested Anatomical Latent embeddings), a clinically grounded framework that learns compact representations from routine non-contrast cardiac CT for major adverse cardiovascular event (MACE) prediction. In 17,659 patients, CARDINAL was evaluated for 1-, 3-, 5-, and 10-year MACE prediction against American Heart Association (AHA) pooled cohort equations (PCE), AHA predicting risk of cardiovascular disease events (PREVENT), coronary artery calcium (CAC), segmentation-derived CT biomarkers, and 70-feature structural radiomics. Gains were largest at longer horizons. At 10 years, CARDINAL (joint) achieved an area under the receiver operating characteristic curve (AUROC) of 0.866 $\pm$ 0.020 and an area under the precision-recall curve (AUPRC) of 0.890 $\pm$ 0.015, compared with an AUROC of 0.826 $\pm$ 0.023 and an AUPRC of 0.826 $\pm$ 0.022 for structural radiomics, the strongest baseline. CARDINAL also achieved the highest survival concordance index (C-index), 0.753 $\pm$ 0.015, and high-versus-low risk-tertile hazard ratio, 10.78 $\pm$ 3.16, with favorable reclassification and exploratory calibration. These findings suggest that non-contrast cardiac CT contains prognostic information beyond conventional risk equations, CAC scoring, and engineered imaging biomarkers.
Read the full abstract
Soft Condensed Matter
Migration of inflated cavities in graded hyperelastic solids
The inflation of a pre-existing, fluid-filled cavity is a timeless topic in the finite-deformation analysis of soft materials. However, classical solutions for cavity inflation rely on radially symmetric material properties, leaving unresolved the effects of non-radial stiffness heterogeneity that are commonly present in biological tissues and engineered soft materials. In this work, we investigate the quasi-static inflation of a pressurized cavity in a hyperelastic solid with shear modulus varying monotonically along a reference Cartesian direction. Finite-element simulations reveal that, beyond an initial small-inflation regime, the most pronounced symmetry-breaking response is the migration of the cavity toward the more compliant end of the material, while nonspherical distortion remains comparatively weak. To analytically quantify this migration-dominated response, we develop a Rayleigh–Ritz reduced-order framework to determine the strain-energy-minimizing migration amplitude for prescribed gradation parameters and inflation level. Without using fitted parameters, the Rayleigh–Ritz framework recovers key features of the cavity migration that are intimately linked to the mechanical gradation parameters. The identification of centroid migration as a salient geometric signal, together with the reduced-order prediction of its evolution, suggests a roadmap for inverse characterization of graded materials through cavity-inflation experiments.
Read the full abstract
Machine Learning
SafeStep: An Interactive Demonstration of Semantic Communication for Pedestrian Safety Monitoring
In this paper, we develop SafeStep, an interactive browser-based semantic communication platform for live pedestrian safety monitoring. SafeStep extracts pedestrian information from four live traffic-camera feeds, transmits it through a semantic communication transceiver over a software-emulated Additive White Gaussian Noise (AWGN) channel, and renders user-specific positions, trajectories, and risk labels. The platform allows each user to select the transceiver, Signal-to-Noise Ratio (SNR), codelength, and Age of Information (AoI) and view the resulting pedestrian reconstruction. SafeStep compares a recently proposed semantic communication design called Meta-VIB with five baseline transceivers. Meta-VIB uses a compact neural model with only $4.16$ million parameters to generalize across varying SNR, codelength, and AoI values without online retraining. Meta-VIB achieves mean task-loss reductions of up to $92.1\%$. On one high-end GPU server, the integrated concurrent-access workload maintains the target $5$ frames/s through $20$ users. At $100$ users, each requesting a distinct configuration, SafeStep records no request failures and a mean application response time below $1$ s, but its mean per-browser frame rate falls to approximately $1$ frame/s. To our knowledge, SafeStep is the first real-time semantic communication platform to make AoI-induced downstream degradation directly observable in live monitoring applications.
Read the full abstract
High Energy Physics (Phenomenology)
Constraining Light and Strange Flavor Equilibration in Relativistic Heavy-Ion Collisions
Relativistic heavy-ion collisions are the only known means of creating the quark-gluon plasma, a deconfined state of QCD matter. Hydrodynamic modeling has established that this medium behaves as a strongly coupled, low-viscosity fluid, but typically assumes that its quark and gluon composition reaches chemical equilibrium by the onset of the hydrodynamic phase. This assumption remains uncertain: gluon-dominated initial states suggest that quark production continues well into the hydrodynamic stage, with potentially different rates for light and strange quarks. This dissertation develops a dynamical framework for light and strange quark chemical equilibration. Incomplete equilibration of each flavor is described through time-dependent quark fugacities that modify both the equation of state and the particlization of the medium. Implemented within a multistage framework of fluctuating initial conditions, viscous hydrodynamics, particlization, and hadronic transport, the model is used to study the effects of equilibration on the hydrodynamic evolution and on hadronic and electromagnetic observables. The framework is embedded within a Bayesian analysis in which Gaussian process emulators trained on the model calculations infer the initial light and strange quark fugacities, their equilibration timescales, and selected transport coefficients from Au+Au collision data at RHIC. The strange sector is found to begin substantially undersaturated, more suppressed than the light sector; the data further favor a slower approach to equilibrium for strangeness, though the timescales themselves remain weakly constrained. These results favor a quark-gluon plasma whose flavor composition is still approaching chemical equilibrium during the hydrodynamic phase rather than fixed at its outset, and motivate extensions to collision systems of differing size and energy.
Read the full abstract
Analysis of PDEs
Incompressible limit for a two-species Brinkman model with drift
We study a two-species model for tissue growth in which both populations are transported by an external drift and by a velocity potential determined through Brinkman's law. The pressure is generated by a stiff constitutive relation depending on the total density. Our main result establishes the incompressible limit as the stiffness exponent tends to infinity, in arbitrary space dimension and for merely integrable initial data. The limit system consists of the two balance laws coupled to Brinkman's equation, the hard-congestion constraint $0\leq n_\infty\leq 1$, the graph relation $p_\infty(1-n_\infty)=0$, and the corresponding complementarity relation. A key point of the analysis is a new $L^2$-based compactness theory that avoids both uniform $L^\infty$-bounds on the pressure and the kinetic reformulation used in earlier approaches. We first construct weak solutions for bounded data and then remove the boundedness assumption by means of a weighted compactness argument inspired by Bresch–Jabin. We also prove an Aubin–Lions–Simon type lemma based on oscillation control, yielding time continuity of the constructed solutions. Finally, a refined dissipation estimate for the Bresch–Jabin compactness functional gives strong compactness of the pressure in the stiff limit and implies a regularising effect: the limiting pressure is bounded even when the approximating pressures are only integrable.
Read the full abstract
Robotics
Distributed Model-Based Diffusion: Finite Horizon Contraction under Bounded Delay
Simultaneously optimizing the trajectories of multiple agents is a challenging problem plagued by nonlinearity, nonconvexity, and the curse of dimensionality. A collection of interacting aerial vehicles or self-driving cars in an intersection are examples of complex multi-agent systems that remain difficult to solve without many simplifying assumptions. The presence of communication latency between agents further increases the difficulty. In this paper, we analyze Distributed Model-Based Diffusion: a sampling-based Model-Predictive Control method suitable for highly nonlinear, nonconvex, nonsmooth, multi-agent systems. We prove contraction and robustness to latency for multi-agent, nonconvex problems, showing applicability to real-world constraints. We test the algorithm on a circleswap task, a cooperative medium-fidelity driving task, and in an aerial combat scenario. Despite the addition of latency, our algorithm improves circleswap makespan by 31% and increases aerial combat win rate by 25% compared to centralized Model-Based Diffusion.
Read the full abstract
Human-Computer Interaction
Deceptive Patterns as a Sociotechnical Phenomenon: Review, Catalog, and Discussion
Background: Deceptive patterns are interface design strategies aimed at misleading users or favoring specific interests, compromising user experiences and ethical privacy principles. These patterns involve exploit different issues of the interaction between humans and technology, revealing gaps in the scientific literature regarding their understanding and working mechanisms. Purpose: This research characterizes deceptive patterns as a sociotechnical phenomenon, integrating human and technical dimensions while offering a self-explainable interactive catalog to raise awareness among users and designers. Methods: To characterize deceptive patterns as a sociotechnical phenomenon, a systematic literature review in Computer Science was conducted, and the Semiotic Framework was applied to analyze and organize the sociotechnical aspects of identified patterns in an integrated manner. An exploratory evaluation of the interactive catalog was conducted, combining heuristic evaluation and focus group. Results: The research identified multiple deceptive patterns in scientific literature. Analysis revealed that existing studies frequently addressed specific aspects of patterns, such as typologies or effects, and do not focus on sociotechnical issues. The exploratory evaluation suggested that the self-explainable catalog is useful and easy to use, with potential to inform users and promote awareness regarding the deceptive patterns existence and functioning. Conclusion: The study contributes to an informed understanding of deceptive patterns, highlighting the need for an approach that integrates human and technical dimensions. The self-explainable catalog is a promising tool to inform and promote awareness about the topic. By exposing the risks and mechanisms of these patterns, the research seeks to promote awareness among users and designers towards ethical interface design practices.
Read the full abstract
Nuclear Experiment
Direct Bayesian Inference of Helicity Amplitudes from Detector-Level Scattering Data
Helicity amplitudes give the most complete description of a variety of scattering reactions used in studies of strong interactions, but they cannot be measured directly: experiments record bilinear combinations of them folded through a detector response, and conventional analyses recover them through multiple stages that introduce discrete ambiguities and require a separate extraction of the absolute cross sections. Here we replace that chain with a single Bayesian inference that determines the experimentally identifiable amplitude parameters directly from detector-level measurements. A score-based diffusion model, trained on a forward simulator, provides the full posterior in each kinematic bin, with the detector response carried by the forward model and positivity of the spin-density matrix guaranteed by the parameterization. The observable amplitude content in electroproduction of final particles grows with polarization of beams and targets, providing additional sensitivity to underlying phases. In simulation, the posterior achieves empirical coverage at or above the nominal level and delivers the angular observables and the separated contributions from longitudinal and transverse photons with their correlations retained. It transfers without retraining to a realistic detector response absent from training and remains reliable in the weakly constrained nucleon-helicity-flip sector, where per-bin likelihood maximization degrades. The result is a general framework for a broad class of inverse problems, phase retrieval, quantum-state tomography, and partial-wave analysis are further instances. Its core requires only a forward simulation of the complete measurement process, instrumental effects are handled within one statistically consistent posterior, applicable across exclusive vector-meson programs at Jefferson Lab, COMPASS, HERMES, and the future Electron-Ion Collider.
Read the full abstract
Quantum Physics
Logical Neural Belief Propagation for Linear-Complexity Decoding of Surface Codes
Quantum error correction (QEC) requires accurate and efficient decoders, yet belief propagation (BP), despite its linear decoding complexity, often provides insufficient logical accuracy on surface codes. We propose Logical Neural Belief Propagation (L-NBP), a BP-based neural decoder that redirects the decoding objective from physical-level to logical-level decoding. L-NBP uses a neural BP (NBP) module to produce posterior beliefs, which a logical classifier transforms into a continuous-valued soft syndrome for logical-operator prediction. Trained end-to-end by backpropagation, the NBP module learns soft syndromes that are favorable for logical classification. On surface codes, L-NBP matches or outperforms BP with ordered-statistics decoding (BP-OSD) and minimum-weight perfect matching (MWPM) while retaining the linear complexity of BP, and achieves a threshold of $17.5\%$ under depolarizing noise. Under circuit-level noise, L-NBP matches the accuracy of BP-OSD on the distance-$9$ surface code while requiring only $0.2\%$ of its complexity.
Read the full abstract
Medical Physics
Physics-Assisted Deep Learning Denoising for Stabilized IMPULSED dMRI Microenvironment Parameter Fitting
Diffusion-weighted MRI (dMRI) is a powerful tool for quantifying cellular microenvironment parameters. This study proposes a physics-assisted deep learning (DL)-based denoising framework designed to enhance dMRI signal quality and improve the robustness of subsequent biophysical model fitting. A dataset of paired noise-free and Rician-noise-corrupted dMRI signals was generated using the IMPULSED-dMRI signal model. Three denoising architectures were evaluated: Convolutional Neural Networks (CNN), Multilayer Perceptron (MLP), and Long Short-Term Memory (LSTM) networks. Denoised signals were then fitted to estimate cell diameter $d$, intracellular volume fraction $V_{\mathrm{in}}$, and extracellular apparent diffusion coefficient $D_\mathrm{ex}$. DL-based processing substantially improved dMRI signal denoising. The MLP and LSTM achieved similar performance, with the LSTM slightly better overall, and both outperformed the CNN. In the subsequent model fitting step, the LSTM produced modest reductions in parameter MAE. The dominant benefit was fitting stabilization, with the overall fitting failure rate reduced from 57.6% to 17.7%. The proposed framework improves dMRI signal quality and stabilizes subsequent IMPULSED-based microenvironmental parameter fitting.
Read the full abstract
Materials Science
Efficient perturbations for basin hopping in amorphous glasses
Efficient exploration of the complex potential-energy landscapes of amorphous materials is central to computational structure discovery and refinement. Conventional Monte Carlo, reverse Monte Carlo, and related methods typically sample configuration space through small, local trial moves and may require millions to tens of millions of moves to converge. Here, we evaluate larger, nonlocal perturbations followed by local geometry relaxation as an alternative sampling strategy. We develop and test four perturbation types using amorphous Al$_2$O$_3$ as a model system. Among them, moving an oxygen atom to change the coordination numbers of two aluminum atoms, allows access to low-energy configurations with substantially fewer trial moves than a conventional Monte Carlo trajectory. These results suggest that relaxation-assisted nonlocal moves could reduce trapping in local minima and improve sampling in structure-search and reverse Monte Carlo workflows.
Read the full abstract
Quantum Physics
High-Dimensional Deterministic Secure Quantum Communication with Reed-Solomon Erasure Coding
Deterministic Secure Quantum Communication (DSQC) is a quantum cryptographic technique engineered to transfer a message through a quantum channel, requiring an auxiliary classical channel for eavesdropping verification and decoding, but without prior key distribution. This article presents a theoretical high-dimensional prepare and measure DSQC protocol using the Reed-Solomon erasure coding to ensure data resilience to noise. This protocol offers the following benefits: it eliminates the need for quantum memory or entanglement, it can be built with commercially available technology, and its higher capacity improves the overall transmission rate.
Read the full abstract
Machine Learning
SegBench-GC: Testing Segmentation Invariance in Multi-Step Offline Goal-Conditioned Reinforcement Learning
Offline goal-conditioned reinforcement learning (GCRL) often uses trajectory structure for future-goal sampling and multi-step targets, yet logged trajectories may be partitioned for administrative reasons that do not correspond to termination. We introduce SegBench-GC, a controlled stress test of segmentation invariance that holds transitions, source trajectories, goal sampling, optimization settings, and evaluation fixed while varying only artificial backup boundaries and whether those boundaries retain continuation value. Continuation-valid targets (CVT) provide the segmentation-consistent control: reward accumulation stops at an artificial cut, but the target bootstraps from its stored successor. In a matched-count PointMaze study with 35,000 artificial cuts, three segmentation realizations, and three optimization seeds, final 50-episode-per-task success is 50.5% uncut, 39.1% with CVT, and 19.1% when the same cuts are treated as absorbing; across segmentation realizations, naive mean success ranges from 4.8% to 31.9%. An independent published n-step baseline (n=25) from the Decoupled Q-Chunking codebase shows the same failure on Puzzle-4x5: 47.2% uncut, 58.5% CVT, and 0.27% naive across three optimization seeds. A target-level diagnostic verifies the analytic target difference to numerical precision, and learned-critic diagnostics show a large optimistic shift under naive handling while CVT remains approximately aligned with the uncut critic. CVT applies standard continuation bootstrapping rather than a new Bellman rule; the contribution is the controlled benchmark, failure isolation, and cross-learner evidence that administrative segmentation can materially change multi-step offline GCRL.
Read the full abstract
High Energy Physics (Phenomenology)
A doubly critical point in the color-superconducting regime of the RG-consistent NJL model
Nambu–Jona-Lasinio (NJL) models of color-superconducting quark matter suffer from cutoff artifacts once temperature or quark chemical potential become comparable to the model cutoff. These artifacts can be removed employing a renormalization-group (RG) consistent regularization scheme. In this article we perform a systematic study of neutral beta-equilibrated quark matter within the three-flavor NJL model with RG consistent regularization. Varying the coupling constant in the scalar diquark channel, we map out the phase diagram in the plane of chemical potential and temperature, and identify the gapless domains. We particularly focus on the melting pattern of the color-flavor locked (CFL) phase. At larger couplings the CFL phase melts through a so-called $d$SC phase, as expected from leading-order Ginzburg–Landau analyses. Lowering the coupling, the phase structure becomes markedly richer: While at large densities the CFL phase still melts through a $dSC$ phase, we find a $uSC$ phase at lower chemical potential. These phases, $uSC$ and $dSC$ meet at a doubly critical point whose existence had been anticipated long ago but was never demonstrated explicitly in a model.
Read the full abstract
Optimization and Control
An Adaptive Augmented Lagrangian Method for Deterministic and Stochastic Nonconvex Optimization
We present an inexact Augmented Lagrangian algorithm for solving nonlinear, non-convex optimization problems. Unlike most recently proposed Augmented Lagrangian methods with worst-case complexity guarantees, we utilize adaptive penalty parameter updates and full dual stepsizes. We show that the method matches the best known worst-case complexity results for Augmented Lagrangian methods (up to logarithmic factors) when both the function and constraints are deterministic, when the function is stochastic and the constraints are deterministic, and when both are stochastic. Experiments on CUTEst test problems confirm the practical advantages of the proposed approach over Augmented Lagrangian methods with non-adaptive penalty parameters and/or short dual step sizes in the deterministic setting. Numerical results on stochastic constrained optimization problems in machine learning also confirm these findings.
Read the full abstract
Artificial Intelligence
Agents for Everyone: A Workshop Framework for Building Agentic AI Capabilities in a Distributed Curation Community
Agentic AI has the potential to accelerate curation of biological databases and knowledge bases. However, uptake has been hindered by a number of challenges and obstacles, including access to agents and appropriate training. Here we describe how we have attempted to address and mitigate these challenges and obstacles through the deployment of a cloud-based agentic environment, and the development of an interactive training workshop for the Gene Ontology Consortium. Our cloud environment for agentic-assisted curation was based on the JupyterHub platform, and utilized Claude Code as a universal harness. This allows curators to interact with an agent session through a terminal running in the browser, and has additional benefits such as centralization of access through a single API gateway, removing the need for participants to manage subscriptions or install software locally. We created four training modules, walking participants through basic agentic tool use first and then working up to agentic biological pathway curation using the existing GO-CAM (GO Causal Activity Model) curation tool. Thirty-seven participants took part in the four-hour workshop. Our key takeaway from this workshop is that building community capability with agentic AI is primarily a problem of access, workflow design, and training. Removing technical barriers, introducing capabilities gradually, grounding exercises in familiar curation tasks, and giving curators direct experience evaluating agent output can provide a practical route toward building shared agentic AI capability in distributed scientific communities.
Read the full abstract
Audio and Speech Processing
Not all generalisation failures can be bought back: four boundaries in affective audio modelling
Models mapping acoustic properties onto affective response underpin applications from music recommendation to sound design, yet are evaluated almost entirely within the corpus they were fitted on. When one fails outside it, the standard response – more data, or a larger model – assumes every failure is a shortage of resources. We show it is not, and that the alternative calls for the opposite remedy. Using four corpora of rated sound, four pretrained representations and three corpora of physiological recording, we pushed one mapping across four boundaries an application must cross: to new material, to edited audio, to a sensor in place of a self-report, and to an individual listener. At each we report the ceiling the target permits, the fraction surviving the crossing, and the price in target-side observations of closing the gap. Within a corpus, prediction reaches 84% of the ceiling set by inter-listener agreement. A same-domain corpus swap costs a fifth of that, and a hundred target labels return two-thirds of the loss. Crossing between music and environmental sound costs four-fifths to all of it, and four pretrained representations recover none of it. Against physiological response no information source we constructed exceeds a third of the attainable ceiling. "The model does not generalise" is therefore two diagnoses, not one, with mutually exclusive remedies; treating the second as the first is the more expensive mistake.
Read the full abstract
Solar and Stellar Astrophysics
A 500 pc volume-limited sample of hot subluminous stars III. The short-period binary population
Hot subdwarf stars of spectral types O and B (sdO/B) in binaries form as products of substantial mass loss near the tip of the red giant branch and offer powerful constraints on binary-star evolutionary models. However, details of some formation channels remain missing. We present a comprehensive analysis of the short-period binary population in the 500 pc volume-limited sample of hot subdwarfs, enabled by precise Gaia DR3 parallaxes. Besides 45 known binaries within 500 pc, this work identifies 50 new single-lined radial-velocity variable systems, 34 with orbital solutions with periods between 0.07 and 22 days. We derive an overall short-period binary fraction of 34.7 (+2.8/-2.9)% for the full sample of 301 hot subdwarfs within 500 pc, corrected for orbital inclination and detection efficiency. The newly solved binaries predominantly occupy the 1-20 day range, a parameter space previously under-represented in the literature. We identify five new reflection-effect systems, three ellipsoidal modulation systems, a newly solved HW Vir system, and a new triple candidate. The sdB and sdO binary fractions are similar, while the sdOB class shows a fraction roughly half that of sdB/sdO stars, with periods over one day, suggesting a different evolutionary pathway. Underluminous hot subdwarfs below the canonical extreme horizontal branch show a binary fraction comparable to other sdB/sdO stars but host only white dwarf companions, which are more massive and mainly found at periods under one day. We provide the first volume-complete birthrate estimates for sub-populations, determining a Galactic merger rate of 2.5+-1.5x10^-5 yr^-1 for sdO/B binaries, explaining no more than ~12% of the eHe-sdO population. We also derive birthrates for two SN Ia progenitor channels, finding hot subdwarf binaries could account for up to 2.5 (+0.7/-0.5)% of the observed Galactic SN Ia rate.
Read the full abstract
Computation and Language
First Make It Playable, Then Make It Good: Staged Interaction Learning for Small Dialogue-Game Agents
We present Qwen-GuidePlay-2B, a 2B-parameter language model for dialogue-game interaction. We fine-tune Qwen3.5-2B using three steps: a) SFT on only successful game trajectories from Playpen, b) weighted turn-level SFT, and c) teacher-guided SFT. The teacher model (which is a larger model) is only used to fix formatting and evaluate examples, but does not create new gold actions. Our final model scores 57.12 clemscore and 42.68 statscore on the public Playpen validation. In the officially released challenge results, our model obtains the second-highest Playpen clemscore delta among submitted systems (which is approximately +36 over its base model). Our findings suggest that imitating full trajectories helps with playability, while turn-level and teacher-guided training usually improve decision-making and increase the overall score. Alternative procedurally heavy approaches like replay-repair and hard-example mining did not help, which suggests that small models are performant simply by using careful curation strategies rather than aggressive changes. We make available both the model and the code for reproducibility.
Read the full abstract
Artificial Intelligence
A Framework for Object-Centric Predictive Monitoring of Collaborative Processes
Predictive Process Monitoring (PPM) of collaborative, inter-organizational processes requires reasoning over multiple interdependent entities, including participants, messages, local executions, and the global collaboration case. Existing approaches extend traditional event logs with collaboration attributes but retain a single-case perspective, leaving much of this structure implicit. Object-centric process mining (OCPM) provides an alternative by representing these entities as first-class objects with explicit relations and multiple notions of case. This study connects collaborative PPM and OCPM through three contributions: (i) a formal semantic mapping from extended collaborative event logs to an OCED-conformant object-centric representation, serialized in OCEL 2.0; (ii) a reformulation of collaborative prediction tasks as object-centric prediction tasks; and (iii) a reproducible converter and prediction pipeline implementing the proposed mapping. We evaluate the framework on four public collaborative event logs and a fifth derived from the BPI Challenge 2013 incident-management log by executing the fourteen reformulated tasks using five predictive strategies across tabular, sequential, and graph-native encodings. We further discuss the benefits, limitations, and threats to the approach's validity. The representation makes collaboration structure explicit and makes it natural to state prediction targets based on object relations that fall outside the case-centric taxonomy, at the cost of increased relational complexity and dependence on object-centric tooling.
Read the full abstract
Neurons and Cognition
A weighted model of perception and decision-making between targets of finite size
Recent research grounded in experiments has connected neural ring models of vision to how animals navigate a complex landscape of attractive targets. In this paper we investigate the mathematical and biological implications of a three-stage model where animals pre-process visual stimuli to identify a discrete set of targets, process this input to select the dominant targets, and then post-process this information to navigate the landscape. Incorporating finite target sizes and a neural density allows consistent target acquisition and pursuit reflecting what is seen in nature. Mathematically, we show this model corresponds to an energy minimization problem, simplifying both its analysis and numerical implementation. Biologically, we argue that the model is physiologically motivated, reproduces experimental observations, and presents a fairly direct pathway to decoding neural geometry via empirical data. Our results also demonstrate an analysis pipeline for future model extensions and suggest specific avenues for experimental research to test and parameterize our quantitative framework.
Read the full abstract
Combinatorics
Majority C-coloring in Cartesian products
A majority C-coloring of a graph $G$ assigns colors to the vertices such that every vertex shares its color with at least half of its neighbors. The maximum number of colors that can be used in such a coloring of $G$ is denoted by $\overline{\chi}_{\geqslant}(G)$. In this paper, the focus is on the majority C-coloring in Cartesian product graphs. It is shown that $\overline{\chi}_{\geqslant}(G \square H) \ge \overline{\chi}_{\geqslant}(G) \overline{\chi}_{\geqslant}(H)$ gives a sharp lower bound, but the difference also can be arbitrarily large. For two-dimensional Hamming graphs, the exact value $\overline{\chi}_{\geqslant}(K_m \square K_n) = \min\{m,n\}$ is established. Balanced Hamming graphs of higher dimension, that is the $k$th powers of complete graphs with respect to the Cartesian product, are also studied. It is proved that $\overline{\chi}_{\geqslant}(K_n^{\square, k})= n^{k/2}$ holds for every even integer $k$. If $k$ is odd and the Hamming graph is the $k$-dimensional hypercube, then $\overline{\chi}_{\geqslant}(K_2^{\square, k})= 2^{\lfloor k/2\rfloor}$. On the other hand, a majority C-coloring of $K_n^{\square, k}$ with at least $3 n^{\lfloor k/2\rfloor}/2 $ colors is presented for every $n \ge 7$ and odd $k \ge 3$. For Cartesian grids, the main result shows that $\overline{\chi}_{\geqslant}(P_m \square P_n) = 1 + \lfloor m/2\rfloor \lfloor n/2\rfloor$ if at least one of $m$ and $n$ is odd, while $\overline{\chi}_{\geqslant}(P_m \square P_n)=mn/4$ holds if both parameters are even and $m \ge n \ge 4$. The paper concludes with a conjecture and several open problems.
Read the full abstract
Computer Vision
Report Supervision
Segmentation models can surpass radiologists, classification models, and vision-language models in tumor detection. Importantly, segmentation models outline tumors, allowing radiologists to better verify and trust the AI output. Their main limitation is the scarcity of tumor masks: creating one 3D tumor mask takes up to 30 minutes, so most public CT datasets contain only a few hundred masks, and even the largest private datasets contain only a couple of thousand. Tumor masks are not produced in clinical routine, but radiology reports are. Public datasets contain tens of thousands of CT-Report pairs, and hospitals contain hundreds of thousands. These reports describe tumors in detail, providing large-scale, informative training data. Here, we introduce Report Supervision (R-Super), a training framework that uses reports to directly supervise and improve tumor segmentation. R-Super introduces new loss functions that teach segmentation models to segment tumors that match report descriptions of tumor count, sizes, and locations. Reports are only used for training. We evaluated R-Super on kidney and pancreatic tumor segmentation, exploring diverse training data sizes, up to 41,418 CT-Report plus 3,488 pancreatic tumor CT-Mask pairs. On external validation, R-Super increased tumor detection F1-Score and segmentation DSC by up to +15% with respect to mask-only training. It also surpassed alternative methods such as CLIP and multi-task learning. Leveraging numerous readily available reports to supplement scarce masks, R-Super strongly improves AI performance when very few training masks are available (e.g., 50), and when many masks are available (e.g., 3,488), unlocking scale in tumor segmentation.
Read the full abstract
Category Theory
A categorical model structure for generalized algebraic theories
We describe a (combinatorial, monoidal, Cat-enriched) Quillen model structure on the category of Cartmell's generalized algebraic theories (gats); its homotopy bicategory consists essentially of Taylor's rooted display map categories. This allows us to compare two kinds of morphisms of gats: one where sort dependency and substitution are preserved strictly, thus directly matching the syntax, and one where the given structure is preserved up to isomorphism. We prove a strictification result for morphisms out of cofibrant theories, which are the retracts of theories without sort equality axioms. Along the way, we give a structural characterization of when a contextual category can be presented without sort equality axioms. Our results also imply that when restricted to cofibrant objects, the tensor product of gats has the expected semantic behaviour, namely, it corresponds to the tensor product of locally finitely presentable categories equipped with a cofibrantly generated weak factorization system. Strict and weak morphisms specialize, respectively, to two familiar concepts of model of a gat A: ones valued in iterated families of sets, with substitution interpreted as reindexing, and set-valued models of the contextual category C(A) viewed as a finite-limit sketch. We characterize strictifiability of a model of the latter kind via a loop freeness condition on a certain map of functors out of the category of context projections of A.
Read the full abstract
Cryptography and Security
Semantic Watermarking with Order-Robust Detection over Sub-sentence Units
Semantic watermarks tie the mark to sentence meaning rather than token choices, promising robustness to content-preserving edits. However, the detector only observes attacker-supplied text, which can be reworded, reordered, or resegmented to evade detection without content loss. Rewording, reordering, and resegmentation all cause embedding displacement: detection tests embeddings different from those selected during watermarking and can therefore lose the mark. Our adaptive embedding displacement attack (EDA) admits all three edits under a single objective that maximizes this displacement. It uses a public paraphraser and surrogate encoder without access to the provider's generator or secret key. At a 5% false-positive rate (FPR) and content-preservation threshold $\bar{q}=90\%$, EDA successfully removes the mark on between 32.6% and 47.9% of documents across four schemes, the highest among the tested attacks. Therefore, EDA evaluates the schemes' robustness more thoroughly than passive paraphrasing. To address these vulnerabilities, we design (k)-SwordStamp: semantic watermarks with order-robust detection over sub-sentence units, reducing sensitivity to attacker-chosen structure at a small quality cost. Against k-SwordStamp, the strongest no-box attack we test is an EDA variant adapted to its design, with a 10.8% attack-success rate. A stronger EDA with access to the provider's detector and secret key reaches a 39.7% attack-success rate, compared with 65.5% on k-SemStamp. Our code is available at https://github.com/D-Diaa/SwordStamp.