Peer Revenue Past citations are no guarantee of future returns

Papers

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

All of arXiv

Friday 9 October

Quantum Physics

Operational signatures of exceptional quantum channels

Paulo A. Brandão

Exceptional points (EPs) are usually introduced as defective degeneracies of non-Hermitian Hamiltonians and, in fully open quantum dynamics, of Liouvillian generators. Here we formulate an operational theory of exceptional points that does not invoke a Hamiltonian or a master equation. The primitive dynamical object is a completely positive trace-preserving (CPTP) quantum channel $\Phi$, applied an integer number $n$ of times, while physical data are probabilities $p_n^{\mathbf A}(x|\rho)$ associated with states $\rho$ and observables $\mathbf A$ having measurements results $x$. We define exceptional channels through nontrivial Jordan chains of $\Phi$ and show that a Jordan block of order $m$ generates discrete polynomial-exponential terms with maximal degree $m-1$. This places the physics of exceptional points in contact with what can be directly measured in experiments.
Read the full abstract

Quantum Physics

Encoding Matters: A Qubit-Efficient Reformulation for Quantum Resilient Routing

Emanuele G. Dalla Torre, Bat-El Friedman and Daniel Porat

Qubits are powerful computational resources because each additional qubit doubles the size of the accessible state space. This expressive power comes at a cost: adding qubits also increases circuit size, constraint overhead, and exposure to hardware noise. These limitations are especially critical for near-term quantum computers, making the number of qubits required to formulate a problem a central practical concern. In this work, we present a method for efficiently mapping logistics and routing problems onto qubits through structure-aware encoding reduction. We demonstrate the method on a benchmark problem of latency-resilient routing under dual-link failures. For this benchmark, the reduced formulation substantially lowers both the qubit count and the compiled circuit complexity while preserving the original optimization task. By running the reduced model on the same trapped-ion quantum computer as the benchmark, we obtain a 100X improvement in the success probability. This finding shows how careful problem formulation can make current quantum hardware substantially more effective for practical optimization.
Read the full abstract

Solar and Stellar Astrophysics

Magnetic connectivity from the Sun to the Earth with MHD models II. Impact of input magnetic maps

Barbara Perri, Silke Kennis, Brigitte Schmieder et al. (+1 more)

Connectivity estimation needs to be fast and thus usually relies on semi-empirical models. We previously established that optimised magnetohydrodynamics (MHD) models can also be used to derive magnetic connectivity, yielding good results, especially for complex configurations. However, the modelling is not the only factor that can impact the connectivity estimation. Models' results, both semi-empirical and MHD, depend heavily on the input data used to drive the simulation. In order to quantify this impact on our connectivity estimation, we used 20 different maps by four providers. We selected the period of July 2019, which corresponds to Carrington rotation (CR) number 2219, because it had already been studied in a previous paper. It corresponds to a period of minimum activity, meaning minimal interference from transient structures, but with a high-speed stream and several polarity switches. We have compared our results with in situ data from ACE and developed several visualisation techniques to validate connectivity not only for a specific date but also for a full CR. The choice of the input magnetic map drastically affects the connectivity estimates, with global scores going from 28% global agreement to 63% in the most extreme case. The HMI and GONG maps consistently gave the best connectivity estimations, with more than 50% correct results of the full CR. For this specific period, the GONG-ADAPT maps yielded the least reliable results, with a score below 35%. This is not surprising, as we have shown before that COCONUT required parameter adjustments for these maps. However, WSO maps which showed poor coronal agreement actually provided good connectivity estimates, showing that the Heliospheric Current Sheet (HCS) location is in the end the most crucial parameter for connectivity. Consistent data over a full Carrington rotation are to be favored compared to daily.
Read the full abstract

All authors Barbara Perri, Silke Kennis, Brigitte Schmieder, Stefaan Poedts

Materials Science

Ferroelectric Domain-Wall Motion, Stability, and Frequency Responses under Different Screening-Charge Conservation Laws

Sergei V. Kalinin

We analyze ferroelectric domain wall motion, shape, and internal structure arising from coupling between polarization and screening charge. Universal cases are organized by electrical boundary conditions, charge conservation and dynamics, and polarization free energy. We consider exposed surfaces, interfaces beneath a dielectric layer and top electrode, and interfaces with a blocking mixed ionic-electronic conductor. Screening occurs through local exchange, exchange with lateral transport, or strictly conserved redistribution. Analytical derivations and finite-depth calculations establish screening drag, velocity branches, and conditions for spatial instability. For an isolated wall joining differently compensated domains, strict surface conservation and zero far-field current exclude steady translation of the screening profile, while allowing transient displacement, finite-sample arrest, and charge-balanced structural changes. Conservation also constrains long-distance transport around a moving junction in an adjoining conductor. Large-field simulations in the top-electrode geometry reveal nonlinear corrugation, transfer toward longer wavelengths, and coupled polarization rotation. First-order polarization energetics allow two interphase fronts enclosing a paraelectric core; multiaxial polarization permits wall rotation and competing intermediate ferroelectric domains. A driven structural-screening reduction supports stable periodic behavior. Reduced-variable diagrams summarize motion and structural regimes, while BaTiO3 and BiFeO3 examples connect them to material thermodynamics. Small-signal analysis distinguishes displacement, polarization, and electrostatic response functions. These results provide a framework for wall dynamics and structural transitions under different driving and screening conditions.
Read the full abstract

Analysis of PDEs

Global Existence of Strong and Weak Solutions to the 2D Compressible Navier-Stokes-Poisson System in Bounded Domains with Large Data and Vacuum

Yingzhi Du and Huijiang Zhao

We study the two-dimensional compressible Navier-Stokes-Poisson system in a general simply connected bounded domain with Navier-slip boundary conditions and a positive, possibly non-flat, doping profile. We establish the global existence of a unique strong solution and of a global weak solution for arbitrarily large initial data allowing vacuum when the shear viscosity is a positive constant and the bulk one proportional to a power of the density with the power bigger than 4/3. The main difficulties arise from the interaction between the Poisson force and the general boundary geometry. To overcome them, we introduce a modified material derivative and derive quantitative $L^p$ estimates for the density by means of a Neumann duality argument together with a harmonic correction generated by the Navier-slip boundary condition. We further obtain a pointwise representation of the effective viscous flux by pulling back the Neumann Green function on the unit disk through a conformal mapping. The resulting commutator estimates, combined with elliptic estimates for the Poisson equation, yield a uniform-in-time upper bound for the density and allow us to close the higher-order a priori estimates. Finally, the global weak solution is constructed through a charge-preserving approximation and compactness argument.
Read the full abstract

Computation and Language

Training User Simulators with Latent Intrinsic Feedback

Tarek Naous, Craig Boutilier, Wei Xu et al. (+1 more)

User simulators are gaining attention as environments to train LLM assistants that can better cope with the nuances of multi-turn conversations with real users. Prior work on user simulation has emphasized generating realistic user utterances, leaving unexplored their potential as a source of direct feedback on the assistant's responses. Such evaluation signals may not be explicitly communicated in language, but can instead reside in the latent reasoning of users. In this work, we introduce user simulators with intrinsic feedback, namely, models equipped with latent evaluation states: internal thoughts, reactions, and scalar ratings that evaluate the assistant from a first-person user perspective at each conversation turn. We train LLM-based simulators using interaction logs between real users and assistants and demonstrate that intrinsic feedback improves simulation realism and critiquing behaviors versus prompted simulator baselines, while also providing turn-level evaluations that existing fine-tuned simulators lack. Further, evaluations generated by our trained simulators correlate better with human judgments than those of prompted simulators of the same size. Lastly, our experiments reveal that simulators with intrinsic feedback provide effective reward signals for multi-turn RL training, matching setups that rely on external judge LLMs for rewards while cutting RL compute by more than half.
Read the full abstract

All authors Tarek Naous, Craig Boutilier, Wei Xu, Ofer Meshi

Signal Processing

Open-H-RF: A Large-Scale Ultrasound Raw Channel Dataset Initiative

T.S.W. Stevens, S.W. Penninga, W.L. van Nierop et al. (+183 more)

Medical ultrasound is a widely accessible, real-time imaging modality, yet progress in data-driven ultrasound research is constrained by the scarcity of large, open collections of raw channel data. Existing public datasets are predominantly based on reconstructed images, while the smaller number of open raw channel datasets is generally specialized to individual reconstruction or imaging tasks, limiting the development of general-purpose methods and foundation models that operate directly on ultrasound signals. While interpretive tasks can be performed on reconstructed images, many of the key technical challenges in ultrasound imaging, such as aberration correction and reverberation reduction, require the raw channel data itself. We introduce Open-H-RF, a community-driven open dataset initiative comprising approximately 40 TB of medical ultrasound raw channel data in 19,332 acquisitions, contributed by 38 institutions across academia and industry, spanning in-vivo, phantom, and simulated data. Each dataset is annotated with a number of supported tasks, ranging from generalized reconstruction, speed-of-sound estimation, and aberration correction to blood flow imaging, motion estimation, quantitative ultrasound, and image interpretation. Open-H-RF provides a unified, self-contained data format that captures raw channel data together with the acquisition, probe, metadata, and reconstruction information required for reproducible processing across scanners and acquisition schemes. The corpus is publicly hosted under a permissive CC BY 4.0 license and can be streamed programmatically and selectively at scale. By bringing heterogeneous ultrasound data into a common representation, backed by modern data infrastructure, Open-H-RF lowers the barrier to working with raw channel data and aims to open medical ultrasound research to new communities. See https://open-h.github.io/Open-H-RF.
Read the full abstract

All authors T.S.W. Stevens, S.W. Penninga, W.L. van Nierop, O. Nolan, F. Duelmer, B.N. Frey, B. Federici, V. van Hal, L.D. van Harten, H. de Hoop, R. Lopata, B. Luijten, M. van Sambeek, V. van der Schaft, H.-M. Schwab, M.F. Azampour, N. Navab, R. Ali, A. Andrzejek, J. Baek, B. Boitnott, T. Brevett, H. Hashemi, D. Hyun, J.H. Park, A.V. Telichko, C. Thomas, L. Zhuang, Y. Aklilu, H. Arora, J. Barker, A.K. Mackanic, P. McDowall, L. McHugh, I. Pirozzi, S. Simon-Trench, A.M. Bronstein, O. Senouf, S. Vedula, D. Zadok, H. Rivaz, M. Sharifzadeh, S. Chen, G.P. Luke, S. Wang, Yujia Wu, J. Verde, R. Göbl, O. Zettinig, E. Boctor, Yixuan Wu, H.-M. Bae, M.-G. Kim, Y.-M. Kim, H. Lee, S.-H. Oh, P. Pinto, M. Berrada-Sounni, M. Carley, R.J. Fedewa, S. Gavin, N. Goldsmith, R. Kassayan, C. Low, R. Mehranfar, G. Bardoscia, M. Boccardo, A. Botter, M. Carbonaro, G.L. Cerone, E. Cesti, K. Meiburger, S. Seoni, F. Truscello, B.J. Raiteri, L. Caron, M. Gasse, N. Ghigo, O. Gulenko, A. Leconte, S.A. Lee, A. Malescot, E. Martineau, J. Porée, J. Provost, G. Ramos-Palacios, B. Rauby, R.L. Rungta, A.F. Sadikot, A. Wu, P. Xing, A. Chavignon, C. Ferlay, V. Hingot, M. Reberol, A. Waelbroeck, H.-S. Kwon, J. Bucholz, I. Guracar, R. Loftman, C. Williams, T. Grutman, T. Ilovitsh, T.A. Aleef, P.B. Chen, W. Chen, Y. Chen, M. Frew, M. Honarvar, Y.C. Lee, Z. Li, H. Moradi, R. Rohling, S.E. Salcudean, Z. Wu, Q. Zeng, P.C. Cattin, S. El Hadramy, N. Bottenus, A.S. Perez, S. Feng, Y. Feng, P. Jin, Y. Lin, O. Ostras, G. Pinton, F. Santibanez, M. Sode, B. Turkbey, H. Wang, B.J. Wood, L. Zhang, S. Luo, J. Wiskin, A. Austeng, Y. Karabiyik, S.P. Näsholm, O.M.H. Rindal, D. Bustamante, M. Mehrmohammadi, H. Raeis, A. Hameddine, G. Lajoinie, H. de Roo, M. Versluis, R. Zorgdrager, J.S. Au, A.J.Y. Chee, J. Cohen, T. Gu, J.Y.-H. Hsu, S. Ling, M. Mourtzakis, H. Nahas, P. de la Torre, D. Xiao, A.C.H. Yu, E. Zafiris, P. Jarosik, P. Karwat, Z. Klimonda, M. Lewandowski, B. Byram, C. Khan, Z. Marshall, Y.-C. Pan, S. Liu, T. Sultan, C. Hoerig, J. Mamou, D. Auerbach, Y.C. Eldar, A. Wegerhoff, G.S. Fischer, S. Katayose, X. Ma, R. Murakami, Y. Tang, R. Tsumura, Y. Wang, H.K. Zhang, M. Azizian, S. Huver, J.J. Dahl, R.J.G. van Sloun, W. Simson

Fluid Dynamics

Predicting the Final Stationary State of Sheared 2D Euler Fluids

Renato Pakter and Yan Levin

The Kelvin-Helmholtz (KH) instability is a fundamental signature of the natural world, manifesting in phenomena ranging from atmospheric clouds to the iconic turbulence of Jupiter's Great Red Spot. While the linear dynamics of this instability is well-understood, the ultimate fate of the fluid — the final stationary state to which it relaxes — remains a formidable challenge. In this paper, we present a general non-equilibrium statistical mechanics framework that allows us to predict the asymptotic vorticity distribution of a sheared 2D Euler fluid. By constructing a self-consistent ansatz for the solution of the Vlasov equation, we determine the stationary stream function without explicitly solving the time-dependent Euler equations. We demonstrate that the relaxation leads to a population-inverted core, where high-energy states are fully occupied, surrounded by a tenuous halo of vortices in the lower-energy regions of configuration space. This result establishes the core-halo structure as a robust statistical consequence of the system's evolution, offering a universal framework for understanding the saturation of shear-induced turbulence in inviscid 2D fluids.
Read the full abstract

Machine Learning (Statistics)

Some remarks on the connection between generative diffusion and optimal transport

Giulio Biroli and Marc Mézard

We discuss three simple connections between generative diffusion and optimal transport. First, the joint law of the data and of their Ornstein-Uhlenbeck noising is, at every time, the solution of an entropic optimal-transport problem, at a temperature set by the noise level. Backward diffusion can consequently be read as an annealing which gradually decreases the entropic regularization. Second, we study the speciation transition–how the modes of the data distribution emerge during generation–and compare how it happens in diffusion and flow matching with respect to optimal transport. Using Gaussian mixtures as high-dimensional example, we show that the dynamical instability through which diffusion and flow matching separate the modes becomes, for optimal transport, an exponentially thin boundary layer at the beginning of generation, which is rounded by entropic regularization. Third, considering a generic forward dynamics with a general confining potential, we show that the backward process driven by the score also satisfies a drift-based variational principle whose minimal cost is the entropy difference between initial and final distributions.
Read the full abstract

Analysis of PDEs

Global Existence and Partial Regularity for the Three-Dimensional Chemotaxis–Navier–Stokes System in a Half-Space

Yubo Chen and Jianbo Yu

We establish endpoint partial regularity for suitable weak solutions to the three-dimensional chemotaxis–Navier–Stokes system with constant sensitivity and linear consumption in a half-space under no-slip and no-flux boundary conditions. The full space–time singular set, including its boundary part, has vanishing one-dimensional parabolic Hausdorff measure. This follows from a Caffarelli–Kohn–Nirenberg type criterion at the natural $r^{-1}$ dissipation scaling, valid at interior and boundary points without additional pressure smallness. The key estimate combines weighted mass control with the localized entropy–energy inequality to close the iteration without a separate $n\log n$ quantity and accounts for the mass term generated by rescaling. Boundary-compatible pressure and caloric decompositions yield local Hölder regularity. We also prove a lower-order criterion requiring smallness only of the velocity, signal gradient, and pressure oscillation, with a threshold depending on an arbitrary finite density mass bound. We construct global suitable weak solutions under natural entropy and energy assumptions without smallness restrictions on the initial data. The potential is time-independent with bounded gradient; boundedness of the potential itself is assumed only for the large-time results. If its gradient is also Hölder continuous, a solution can be selected to become bounded and classical throughout the closed half-space after a finite waiting time. For the same solution, we obtain decay and Neumann Gaussian asymptotics for both scalar components without initial spatial moment assumptions. The profile coefficients are the conserved bacterial mass and the limiting signal mass, respectively, and the latter is strictly positive whenever the initial signal is nonzero.
Read the full abstract

High Energy Physics (Theory)

The Gregory-Laflamme Instability of five-dimensional Black Strings in Einstein-Gauss-Bonnet Gravity

Pau Figueras, Áron D. Kovács and Shunhui Yao

We complement and extend our recent study of the non-linear evolution of the Gregory-Laflamme instability of black strings in Einstein-Gauss-Bonnet gravity in five spacetime dimensions. First, we study time-dependent perturbations of uniform black strings in Einstein-Gauss-Bonnet theory and determine the growth rates as a function of the coupling. We find that for positive Gauss-Bonnet couplings, the black strings are more unstable than in general relativity in the sense that the growth rates are larger and shorter wavelengths are unstable. In the second part of the paper we show that the hyperbolicity of the equation of motion of Einstein-Gauss-Bonnet theory is determined by an effective metric when the dynamics is restricted to spherical symmetry in the asymptotically flat directions. When the effective field theory becomes strongly coupled, as it happens near the endpoint of the Gregory-Laflamme instability of the black string, this effective metric becomes degenerate and the Cauchy evolution breaks down. Applying this criterion to uniform black strings, we find that for any positive Gauss-Bonnet coupling there is an open set of regular solutions (including the one of minimal thickness) on which the effective metric degenerates in a neighbourhood of the horizon. Therefore, we argue that these solutions cannot be attained from weakly coupled initial data by Cauchy evolution and the solutions of the linearized equations around them cannot be interpreted as approximations to nearby non-linear evolutions. We provide further evidence for this picture with non-linear simulations of the Gregory-Laflamme instability. The non-linear simulations terminate through a loss of hyperbolicity of the Tricomi or Keldysh type, depending on the sign of the Gauss-Bonnet coupling.
Read the full abstract

Social and Information Networks

Non-Bayesian Social Learning with Heterogeneous and Time-Varying Aggregation Rules

Sammy Oukaci, Pol Rodríguez Farrés and Jesus Cerquides

People constantly form opinions by combining what they observe with what friends, collegues, and communities tell them. Following this idea, this article generalizes earlier results in Non-Bayesian social learning on strongly connected social networks, where agents repeatedly update their beliefs by combining their own private signals with the beliefs of their neighbors. Previous work assumed that every agent aggregates information in the same way and often does not change in time, assuming agents never change how they reason and listen to others. We show that agents' beliefs still converge almost surely to the true state when agents use heterogeneous and time-varying (non-stochastic) aggregation rules under certain conditions. That is, even when people change their way to aggregate others opinions, the community can still reach the truth as long as information flows among its members. We also provide convergence rate results under this scenario and show that, for some of these functions, it is at least equal to the best convergence rate we have found in the literature. We present the widely used aggregation functions in the Constant Elasticity of Substitution (CES) family as a specific case of our general model, which include rules such as the log-linear, DeGroot, and the harmonic rules, and provide a simple illustrative example to show the capabilities of the model.
Read the full abstract

Computer Vision

CUBE: Class-conditioned Unknown Boundary Evidence for Open-Set Insect Species Recognition

Hao Tan, Qimin Zhang, Zhe Xiao et al. (+3 more)

Open-set insect species recognition faces significant challenges when deployed in real-world biodiversity scenarios, particularly when encountering unknown species that are visually close to known training classes. Unlike generic out-of-distribution samples, fine-grained unknown insects are often not visually anomalous. They instead tend to appear near the decision boundary of specific known species, leading to unreliable closed-set predictions. To address these challenges, we propose Class-conditioned Unknown Boundary Evidence (CUBE), a novel evidence-based framework that models unknownness around each known species rather than as a single global unknown class. Our key idea is to evaluate a test sample relative to the known species it most closely resembles, enabling local rejection for visually confusable unknown insects. CUBE learns class-conditioned boundary evidence from auxiliary species, preserves auxiliary species as shadow references during inference, and incorporates taxonomic candidate calibration to verify biological plausibility. When ecological metadata is available, we further introduce CUBE-E, which adds a lightweight location-consistency term without modifying the visual backbone. Experiments on the Open-Insect C-America benchmark show that CUBE achieves superior performance, outperforming state-of-the-art baselines with 95.8%, 98.7%, and 99.1% AUROC on local, non-local, and non-moth unknown splits, respectively, while maintaining 94.9% closed-set accuracy. With ecological metadata, CUBE-E further improves the results to 96.0%, 99.7%, and 99.5% AUROC with 95.2% closed-set accuracy.
Read the full abstract

All authors Hao Tan, Qimin Zhang, Zhe Xiao, Zekai Zuo, Wanting Liu, Jinghui Qin

Quantum Physics

Genuine multipartite nonlocality detection in graph states

Owidiusz Makuta, Flavio Baccari and Remigiusz Augusiak

Experimental detection of multipartite Bell nonlocality faces many challenges. The increase in the size of the system demands a higher fidelity of state preparation, a higher detection efficiency, and a larger number of correlators that need to be estimated. These issues are even more pronounced if one wishes to detect a more restrictive variant of Bell nonlocality called genuine multipartite nonlocality (GMNL). In this work, we introduce two classes of multipartite Bell inequalities that are capable of detecting GMNL correlations and address different experimental hurdles in GMNL detection. The first class consists of the number of correlators that scales linearly with the number of parties. The second class increases the number of correlators and, in return, provides higher noise robustness of GMNL detection. Both classes of inequalities are constructed to detect GMNL in graph states, and in fact both can be used for self-testing of graph states and the performed measurements.
Read the full abstract

Audio and Speech Processing

Vocal Intensity Conversion with Continuous SPL Control via Adversarial Training

Quentin Le Tellier, Albert Rilliard, Olivier Perrotin et al. (+1 more)

Vocal intensity conversion renders a recorded utterance at a different vocal intensity, rather than simply scaling its amplitude, while preserving its content and speaker identity. Existing systems control intensity through a reference utterance or a few discrete categories. Instead, we control the conversion with any target sound pressure level (SPL) on a continuous dB scale. A lightweight converter, trained as a conditional cycle-consistent GAN without parallel data, transforms the features of a frozen speech encoder, from which a pretrained vocoder resynthesizes the speech. We compare three representations, log-mel spectrograms, WavLM, and SpeechTokenizer, and introduce an objective protocol that measures the achieved intensity change against the target one on a calibrated corpus. With WavLM, the converter achieves a measured 83% of the target change in intensity and also shifts the fundamental frequency. A perceptual evaluation confirms that perceived vocal intensity increases with the target level on amplitude-normalized stimuli.
Read the full abstract

All authors Quentin Le Tellier, Albert Rilliard, Olivier Perrotin, Marc Evrard

2610.13608 · 9 Oct · 5 pages, 2 figures, 1 table. Submitted to IEEE ICASSP 2027. Code: https://github.com/vers-project/vocal-intensity-conversion. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

Abstract PDF $SPCH ×2$ADVR ×1

Computation and Language

BLiMPPI: Benchmark of Linguistic Minimal Pairs for Polarity Items

Arzu Burcu Güven, Anna Rogers, Leonie Weissweiler et al. (+2 more)

Negative polarity items (NPIs), such as "anymore" and "ever", are grammatical only in select environments, predominantly in the context of negation. Because their grammaticality depends on both syntax and semantics, they have a rich, multilingual literature in theoretical linguistics, yet work on NPIs in language models is almost entirely confined to English. We introduce BLiMPPI, a multilingual evaluation suite of over 12,000 NPI minimal pairs in 33 languages, built by a pipeline that discovers NPIs in Universal Dependencies (UD) treebanks, labels their licensing environments, and perturbs these into minimal pairs. Evaluating 12 multilingual and 2 monolingual models alongside n-gram baselines, we find that multilingual models struggle most with lower-resource languages and with pairs that are not n-gram solvable. We further examine the effects of NPI typology, language family, licensing variability, and NPI-licenser distance. Data: https://github.com/arzuburcuguven/blimppi-data
Read the full abstract

All authors Arzu Burcu Güven, Anna Rogers, Leonie Weissweiler, Joakim Nivre, Rob van der Goot

Machine Learning (Statistics)

CorrFlow: Exact and Permutation-Equivariant Flow Matching on the Correlation Elliptope

Gautier Marti

Correlation matrices combine nonlinear positive-definiteness constraints with a simultaneous row-column permutation symmetry. Their generation therefore involves two distinct problems. First, every output must remain on the correlation elliptope, irrespective of how its variables are ordered. Second, the generator must learn the domain-specific distribution occupied by empirical data inside that valid space. Existing financial image generators mix these problems: they learn ordered matrix entries and repair invalid draws afterward. We introduce CorrFlow, which separates them. A generalized Fisher transform maps full-rank correlation matrices bijectively to unconstrained coordinates, guaranteeing valid outputs without projection. A graph flow-matching network learns the empirical distribution in those coordinates and is permutation equivariant: reordering the variables only reorders the output. We compare a conditional CorrGAN reimplementation, image diffusion, and CorrFlow on $80\times80$ Taiwan equity matrices using five seeds per neural method. CorrFlow produces 100% valid draws without repair and obtains feature sliced-Wasserstein distance $.712\pm.141$, versus $4.339\pm.252$ for CorrGAN and $4.610\pm.191$ for image diffusion after their invalid outputs are projected back to the correlation space. It also improves this distance over historical lookup, while lookup retains higher local coverage. These results support the combined construction under the evaluated training budget; they do not establish superiority over fully tuned image models or isolate the contribution of each component.
Read the full abstract

Computation and Language

zip2zip++: Towards Inference-Time Token Compression with Minimal Quality Loss

Saibo Geng, Andrea Gentilini, Xinxian Ma et al. (+3 more)

Dynamic vocabularies built on-the-fly via Lempel-Ziv-Welch (LZW) compression let a language model represent the same text in fewer tokens, but prior work shows that this efficiency comes at a cost to reasoning and perplexity. We show that this quality gap is not a fundamental price of compression but can be substantially closed through targeted improvements to the compression layer and LZW codec, without modifying the underlying pretrained language model. Our analysis reveals four key issues: (1) LZW's compression algorithm creates an implicit conflict between the input embedding and the output embedding, a tension that a single shared embedding cannot resolve; (2) naively applying RoPE to LZW-compressed sequences ignores that merged tokens can span multiple tokens; (3) merging certain tokens (notably digits) disproportionately hurts downstream accuracy; and (4) dynamic vocabularies introduce an inherent segmentation ambiguity that confounds standard evaluation. We address these through decoupled input/output dynamic embedding layer, size-aware RoPE positioning, and a modified LZW codec supporting selective merges and multi-view segmentations. Across four models, our recipe improves perplexity on every corpus we evaluated as well as mathematical reasoning accuracy, closing roughly half the quality gap between zip2zip and the uncompressed baseline at a similar compression ratio on general text. This represents a significant step toward lossless adaptive-vocabulary compression for language models.
Read the full abstract

All authors Saibo Geng, Andrea Gentilini, Xinxian Ma, Nathan Ranchin, Yunzhen Yao, Robert West

Computer Vision

Thyroid Nodule Segmentation in Ultrasound Images with Conformal Uncertainty Estimation

Francesco Teodori, Pierpaolo Brutti and Umberto Ferraro Petrillo

Thyroid nodules affect a large proportion of the general population. Although most are benign, a non-negligible fraction is malignant. Ultrasound imaging is the primary diagnostic modality, and accurate nodule segmentation is a prerequisite for CAD systems to classify malignancy and assess metastatic risk.\\ In this work, we propose a thyroid nodule segmentation framework that integrates a DeepLabV3+, channel-attention modules and post-hoc conformal prediction, providing both accurate mask and pixel-level uncertainty estimates. We address thyroid ultrasound nodule segmentation as the first stage of a pipeline that also includes malignancy classification and metastatic risk prediction. We adopt DeepLabV3+, whose Atrous Spatial Pyramid Pooling (ASPP) captures multi-scale context, with an EfficientNet-B5 backbone for its accuracy-efficiency trade-off. Additionally, we test three channel-attention modules: Squeeze-and-Excitation, CBAM and Grouped Channel Attention (GCA), paired with two loss formulations (Dice-BCE and Boundary).\\ Since binary masks do not indicate prediction confidence, we wrap the trained model with a split conformal prediction layer that produces pixel-level uncertainty maps with statistical coverage guarantees.\\ We evaluate all configurations on the TN3K public benchmark using a fixed split and directly compare them with state-of-the-art methods. The plain backbone already improves upon previously reported results, with CBAM + Dice-BCE emerging as the best overall configuration.
Read the full abstract

Machine Learning

A small cognitive core explains human behavior at scale

Marcel Binz and Eric Schulz

Psychology has proposed many principles of cognition. Each, however, was developed in isolation, leaving their combined explanatory power unknown. We assembled a small set of established principles into one model, which we call COntextual Retrieval and Encoding (CORE). CORE consists of a hierarchy of episodic memories into which prediction errors are written and whose encoding, retrieval, and forgetting depend on context. Fitted jointly to 7,036,442 choices from 137 experiments, the model predicts human choices better than black-box models and reproduces 75% of reported behavioral effects when simulated. Moreover, its latent variables show classic signatures of Bayesian learning and event segmentation. Taken together, these findings suggest that a small cognitive core accounts for much of human behavior. They also open a path toward making psychology more than the sum of its parts, by building computational models on the entire experimental record.
Read the full abstract

Quantum Physics

Entanglement Catalysis Beyond the Asymptotic Rate

Alexander Streltsov

Quantum entanglement is a fundamental resource in quantum information science, commonly consumed to perform tasks like quantum teleportation and quantum key distribution. Entanglement concentration converts partially entangled pure states into nearly perfect Bell pairs. While its efficiency is well understood in the asymptotic limit, practical protocols must operate with finite systems. In this finite regime, and at the low error levels relevant in practice, performance is penalized by corrections that reduce the achievable yield. We investigate whether this fundamental finite-size penalty can be mitigated by an entanglement catalyst: an auxiliary state that aids the transformation but is returned exactly and uncorrupted. Although such a catalyst cannot increase the leading-order asymptotic yield, we show that it fundamentally alters the finite-size scaling. For every bipartite pure state with nonuniform Schmidt coefficients, we exactly determine the optimal second-order correction to the yield and show that catalysis completely overcomes this finite-size penalty.
Read the full abstract

Spectral Theory

Non-resonant angles in a Robin eigenvalue problem

Konstantin Pankrashkin

In an earlier study of the Laplacian eigenvalues with strongly attractive Robin boundary conditions in polygons it was found that if each corner is either concave or non-resonant, then the asymptotic behavior of the individual eigenvalues is determined by an effective one-dimensional operator on the boundary with Dirichlet boundary conditions at the corners. The non-resonance condition was formulated in terms of an eigenvalue bound on truncated sectors, and it was only shown to be satisfied by the right and obtuse opening angles. In the present work, we determine the full range of non-resonant angles: it turns out that the non-resonance condition fails only for the opening angles $\frac{\pi}{2n+1}$ with $n=1,2,3,\dots$, and these critical values are equivalently characterized in terms of the jumps of the eigenvalue counting function and the existence of bounded generalized eigenfunctions. This resolves the numerically supported conjecture stated by Dauge and Lafranche.
Read the full abstract

Sound

AudioTETRIS: Token Elimination via Transitional Importance Signals for Efficient Audio LLMs

Shreya Sajid, Anna Vettoruzzo and Joaquin Vanschoren

Large Audio Language Models (LALMs) achieve strong performance on speech-text tasks, but standard audio encoders produce hundreds of tokens for even short audio clips. Because self-attention scales quadratically in sequence length, this inflates inference-time compute substantially. Current training-free token compression methods prune or merge tokens based solely on coarse attention or feature similarity. However, treating audio tokens as generic, isotropic features smooths over spectral transitions, destroying fine-grained phonetic transitions essential for decoding speech. In contrast, speech signals exhibit highly structured temporal patterns where information density peaks at acoustic transitions. We propose AudioTETRIS, a training-free, two-stage framework that uses the audio encoder's representation geometry as an acoustic prior for token importance before applying attention-based compression. Instead of discarding encoder structure after the encoding stage, AudioTETRIS propagates this prior directly into the LLM decoder to compute a per-token fused score. This score drives progressive top-$k$ token selection across decoder layers without requiring additional training or auxiliary models. Evaluated on Qwen2-Audio-7B, Phi-4-Multimodal, and Audio Flamingo 3, AudioTETRIS retains near-uncompressed performance on several speech tasks using only 25% of audio tokens, matches or surpasses full-token accuracy at moderate compression rates, and consistently outperforms existing training-free baselines.
Read the full abstract

Machine Learning

EchoTab: Learning Measurement-Grounded Representations for Echocardiographic Diagnosis through Tabular Reconstruction

Michelle Espranita Liman, Ioana Ionac, Alexander Müller et al. (+3 more)

Diagnostic information in multi-view medical imaging studies, like echocardiography, is distributed across views, each capturing different aspects of cardiac structure and function. In clinical practice, these views are first interpreted through quantitative measurements, which then support specific diagnoses. However, existing echo representation learning methods do not explicitly model measurements as an intermediate representation between views and diagnosis. As a result, individual measurements are not explicitly represented in a form that downstream diagnostic tasks can selectively use, hindering accurate diagnosis. Inspired by this hierarchical clinical workflow, we introduce EchoTab, a measurement-grounded representation learning framework that represents multi-view echocardiography studies using measurement-specific tokens as an intermediate representation between echo views and diagnosis. Each measurement token selectively attends to the views relevant to its estimation and is trained through tabular reconstruction to predict the corresponding measurement value. The resulting tokens form a factorized echo representation, which encode clinically meaningful structural and functional properties and can be selectively used for downstream diagnosis. Unlike the commonly used contrastive approaches, EchoTab does not require negative samples or large batch sizes. Experiments on two datasets demonstrate that EchoTab learns effective measurement-grounded representations that improve diagnosis classification in multi-view echocardiography and its learned attention aligns with established measurement-view relationships used by cardiologists. Our code is accessible at https://github.com/michelleespranita/EchoTab.
Read the full abstract

All authors Michelle Espranita Liman, Ioana Ionac, Alexander Müller, Eimo Martens, Daniel Rueckert, Philip Müller

Statistics Theory

Beyond Discrepancy: Oracle Drift in Nonstationary Weighted Risk Minimization

Tobias Brock and Thomas Nagler

We analyze distributional drift in nonstationary weighted risk minimization by decomposing the excess risk into estimation, drift, and approximation components. The drift error is commonly bounded through discrepancy-type quantities that control changes of the risk over classes of predictors. We instead relate the drift error directly to the movement of the best predictor within the model class and analyze its behavior in parametrized regression models. Our analysis shows that discrepancy-type quantities can be overly conservative: unlike our oracle drift, they may scale at a larger order than the resulting excess-risk contribution and can respond to changes in the risk surface that do not affect the optimal predictor. For linear regression, we characterize this gap sharply and combine the resulting drift bounds with estimation guarantees. We further extend the analysis to nonparametric regression.
Read the full abstract

Cryptography and Security

A Systematic Study of Security Enforcement at Android Inter-Component Boundaries

Sandiyo Christan A

Android's inter-process communication (IPC) mechanisms (Activities, Services, BroadcastReceivers, ContentProviders, and URI scheme handlers) form the primary interface between applications and between applications and the system. We present a systematic study of security enforcement comprising 120 evaluations (99 inter-component interface evaluations across five interface categories, 10 platform security-control evaluations, and 11 dedicated positive controls) in over 30 Google and AOSP packages on Android 17 (API 37). Using an expected-versus-observed framework under a zero-permission attacker model, we evaluate whether each boundary's enforcement mechanism functions as designed. We find 97 enforcement gaps (80.8%) and 23 enforcement holds (19.2%). From these observations, we derive a 14-category enforcement gap taxonomy organized across manifest, code, platform, and architectural layers. Within our purposively selected corpus, observed enforcement outcomes differ substantially across interface categories: evaluated Activities and BroadcastReceivers exhibit a 100% gap rate, while ContentProviders, which benefit from framework-provided defenses such as setStrict(), achieve a 27.6% holds rate. We find that all seven evaluated platform compatibility changes with security intent are either disabled or gated at a targetSdkVersion that none of the applications in our evaluated corpus targets. We document 18 custom URI scheme registrations handling sensitive data (authentication tokens, payment credentials, medical information) despite the known unverifiability of custom schemes (RFC 8252 Section 8.1). Our study includes 23 positive controls, calibration cases demonstrating that the evaluation procedure distinguishes functioning enforcement from observed enforcement failures.
Read the full abstract

Functional Analysis

Gaussian inequalities with different covariance matrices - the Heterovariant Gaussian Brunn-Minkowski inequality

Emanuel Milman, Shohei Nakamura and Hiroshi Tsuji

We establish the following novel heterovariant Gaussian Brunn-Minkowski inequality (involving general covariance matrices $\Sigma_1,\Sigma_2,\Gamma$) for origin-symmetric convex sets $K,L \subset \mathbb{R}^n$ and $p,\lambda \in (0,1)$: \[ \gamma_{\Sigma_1}(K)^{1-p} \gamma_{\Sigma_2}(L)^p \le \left( \left( \frac{1-p}{1-\lambda} \right)^{1-p} \left( \frac{p}{\lambda} \right)^p \right)^{n} \gamma_{\Gamma}((1-\lambda) K + \lambda L) , \] whenever $\Gamma \leq \Sigma_1 \sharp_p \Sigma_2$, the $p$-geometric mean of $\Sigma_1,\Sigma_2$. This immediately implies both the dimensional Gaussian Brunn-Minkowski inequality of Eskenazis-Moschidis (by using $\Sigma_1 = \Sigma_2 = \Gamma$ and optimizing over $p$) and the strong Gaussian (B)-theorem of Cordero-Erausquin-Fradelizi-Maurey (by using $K=L$, $p=\lambda$ and diagonal covariances), fusing them into a single unified statement. In fact, the inequality applies to convex sets whose appropriate Gaussian barycenter is at the origin. Furthermore, it reveals a new *bilinear* version of the strong Gaussian (B)-theorem (involving *two* convex sets); in particular, for all $a,b > 0$ and $\lambda \in [0,1]$: \[ \gamma(a K)^{1-\lambda} \gamma(b L)^{\lambda} \leq \gamma \left ( a^{1-\lambda} b^\lambda \left ((1-\lambda) K + \lambda L \right) \right ) . \] A more general heterovariant version of the (bilinear) (B)-theorem replaces dilation by general linear transformations. A different version yields a heterovariant form of the recently established Gaussian conjugate Rogers-Shephard inequality. By averaging these Gaussian inequalities with respect to the covariance matrices, we obtain dimensional Brunn-Minkowski and conjugate Rogers-Shephard inequalities for Gaussian mixtures. The proofs are based on the Gaussian Forward-Reverse Brascamp-Lieb inequality established in our previous work.
Read the full abstract

Physics and Society

A data-driven heuristic-ensemble method for behavioural inference in collective-risk dilemma experiments

Marc Sadurní, Martín F. Díaz, Josep Perelló et al. (+1 more)

The collective-risk dilemma has become a paradigm for studying cooperation in a wide variety of social experiments, initiatives that generate increasingly rich datasets describing individual actions over successive decision-making processes. Beyond the purely descriptive analysis of such empirical datasets, a methodological framework is required to interpret and reproduce the observed dynamics, a scheme that may be skewed by exogenous biases. Here, we introduce a data-driven, non-parametric method for simulating individual decisions together with a validation protocol for assessing the fidelity level of the outcomes they generate. Our methodological contribution is based on an empirical heuristic ensemble: agents are represented by a finite set of behavioural heuristics, probabilities conditioned on different observable features of the game state, endowed with a hard-voting arbitration that determines their per-round final decisions. This reproducible generative framework allows for behavioural inference in different collective-risk dilemma setups: Behavioural heuristics are seeded directly from empirical data assessing conditional probabilities, without a priori assumptions; a plurality vote with uniform random tie-breaking resolves between competing heuristics each round, adding no free parameters; validation follows a pattern-oriented approach, comparing synthetic and empirical macro-indicators against single-heuristic and uniform-random baselines; synthetic data generated allow exploration of scenarios or experimental setups that have not been empirically realised; as a worked example, we instantiate our methodological approach on the xAire citizen science collective-risk dilemma experiment run in Barcelona.
Read the full abstract

All authors Marc Sadurní, Martín F. Díaz, Josep Perelló, Miquel Montero

Artificial Intelligence

MENTOR: Learning from Failure with On-Policy Pedagogical Intervention

Guanlan Yue, Hang Yu, Xiaoyong Huang et al. (+4 more)

Reinforcement Learning from Verifiable Rewards (RLVR) is a key driver for Large Reasoning Models, but it is vulnerable to "zero-reward collapse", a training failure where a sudden increase in task difficulty causes the learning signal to vanish. We argue that the conventional approach of discarding these failed trajectories is wasteful, as they contain valuable diagnostic information about the model's cognitive gaps, a bright side that can be leveraged for recovery. Existing solutions, however, often regress to static, "off-policy-style" interventions like pre-defined hint libraries or perform post-hoc refinement, failing to provide the adaptive, forward-looking guidance needed to escape the collapse. To address this, we propose a paradigm shift towards on-policy pedagogical intervention and introduce MENTOR, a multi-agent framework that acts as a live mentor. MENTOR diagnoses the root cause of failure in real-time and dynamically generates targeted, instructional prompts to guide the learning agent's next attempt. As a plug-and-play module, MENTOR integrates seamlessly into existing policy optimization pipelines like GRPO without modifying the training infrastructure. Extensive experiments on math and code generation show that MENTOR effectively mitigates zero-reward collapse, outperforming state-of-the-art baselines. Notably, it achieves massive average relative gains of up to 78.3% on Llama-3.2-3B-Instruct over its base model.
Read the full abstract

All authors Guanlan Yue, Hang Yu, Xiaoyong Huang, Ruipeng Li, Wei Zhang, Bingchang Liu, Rui Wang

Databases

Time-Varying Favorite-Longshot Bias in Polymarket Weather Markets

Shouchen Qiu and Huaijun Qiu

Prediction-market prices combine traders' judgments about future outcomes, but their errors may be concentrated in particular price ranges and change as settlement approaches. Prior cross-category evidence shows that Polymarket weather markets have larger pricing bias than other major categories, without separating low-price longshots from high-price favorites. This study compares the two ranges at different horizons in daily maximum-temperature markets. Event-level probability scores, cubic B-spline recalibration, and Buy-NO tests are used to examine the form and effect of the bias. Longshot overpricing mainly occurs at distant horizons, whereas favorite overpricing becomes much stronger during the final 1-12 hours and is nearly one order of magnitude larger. Event-level scores improve toward the daily temperature peak, while simple recalibration yields no statistically significant gains. The Buy-NO results in the corresponding time windows also show that the main pricing error is concentrated at the favorite end.
Read the full abstract