Peer Revenue Now accepting positions

Papers

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

All of arXiv

Friday 6 November

Systems and Control

An Optimal Power-Flow Grouping Topology for Single-Isolation-Stage Multi-Output Solid-State Transformers

Electric-vehicle dc fast-charging stations and data centers require multiple high-power dc outputs with independently varying demands. Conventional multi-output solid-state transformers typically establish a common dc bus and use downstream dc/dc converters for port-level regulation and, where required, galvanic isolation. A single-isolation-stage solid-state transformer can eliminate this additional high-power interface stage by directly forming multiple isolated outputs from its submodule dc/dc stages. However, because one submodule is generally insufficient for an individual output port, multiple submodules must be connected in parallel, while inter-submodule balancing circuits redistribute power under unbalanced loading. The submodule grouping therefore determines the internal balancing-power paths. Sequential grouping can create long transfer paths and accumulated balancing power, increasing current stress and losses. This paper proposes an optimal power-flow grouping topology that interleaves submodules assigned to different output ports, thereby localizing power redistribution and reducing the power processed by the balancing circuits. Analytical models quantify the required balancing power and its reduction relative to sequential grouping. A coordinated submodule-bypass strategy is also developed, and the remaining output-power capability under different fault scenarios is evaluated. A six-submodule prototype validates the proposed topology. Under the same highly unbalanced operating conditions, the proposed topology reduces the maximum submodule dc-link voltage difference by 63% and the peak balancing current by 14.5%, while improving the measured efficiency by approximately 5.6% at rated power. Stable post-fault operation is also experimentally demonstrated.
Read the full abstract

All authors Yi Hao, Krzysztof Puczko, Levy Ferreira Costa, Liangcai Shu, Dongsheng Yang

Computer Vision

T2I Models Rarely Invert Negation but Often Fail to Suppress Negated Concepts: Benchmarking Negation Suppression in Text-to-Image Generation

Text-to-image diffusion models often fail to suppress explicitly negated concepts, such as a lamp for "a desk without a lamp." We introduce T2I-NSI, a benchmark and metric for measuring negation sensitivity in text-to-image generation. It pairs each negated prompt with a canonical positive counterpart and factorizes negation by linguistic form (lexical/semantic, morphological, syntactic) and target type (objects, attributes, relations, compositions, and nested cases). The benchmark contains 759,869 paired prompt atoms over a 5,082-object catalog, with 916,272 images (911,982 open-weights, 4,290 closed-source), scored by a Basic CLIP estimator and a Robust ensemble of CLIP, VQA and open-vocabulary detection. Negation sensitivity is low: open-weights generators score NSI <= 0.038, while GPT-Image-1.5 and Gemini-2.5-Flash-Image reach about 0.09. The dominant failure is under-suppression rather than misdirection. Inverse interpretation is essentially absent (0.00% across over 450K atom-model evaluations), whereas models render the concept under both prompts on 8-19% of atoms and only partly suppress it on a further 9-25%; most remaining atoms are too weakly rendered to label. A blinded human study (420 atoms, nine annotators) corroborates the near-absence of inversion and the high persistence rate, and a matched-prompt comparison shows the open-versus-closed gap (~3.5x) is not a prompt-generator artifact. Code and data: https://github.com/IAB-IITJ/T2I-NSI
Read the full abstract

All authors Akansha Gautam, Aparna Bharati, Mayank Vatsa, Richa Singh

Differential Geometry

Natural potentials for TT tensors, divergence-free symmetric tensors and linearised relativistic initial data

Let $N$ be a natural tensor which vanishes at a metric $g\_0$. By diffeomorphism invariance, $DN(g\_0)^*$ takes values in divergence-free symmetric $2$-tensors, and in TT tensors if moreover $DN(g\_0)$ kills conformal directions. Such operators are local potentials, which we call machines: they produce smooth compactly supported tensors with the required property and, on closed manifolds, all of them up to a finite-dimensional space when their symbol is injective. On Einstein manifolds of every dimension, the linearised Bach tensor gives the operator $P$ of an earlier paper, which factorises on TT tensors into the massless and partially massless spin-$2$ operators; we also construct machines on Bach-flat, conformally flat and Ricci-parallel metrics. Replacing the diffeomorphisms of $M$ by those of the vacuum development, we obtain potentials, with a Fredholm property, for the linearised Einstein constraint equations on arbitrary spacelike hypersurfaces of Minkowski, de Sitter and anti-de Sitter spacetimes.
Read the full abstract

Machine Learning

A Stochastic Latent Variable Generalization of Joint-Embedding Predictive Architectures

Joint-Embedding Predictive Architectures (JEPAs) predict the latent embedding of a target signal from a context signal, with an auxiliary variable z conditioning the predictor. Despite its prominent role in the original JEPA blueprint, the content of z has been carried by a fixed deterministic token in every prior instantiation, collapsing the genuinely multi-modal distribution over plausible completions to a point estimate. We present S-JEPA, which realises z as a proper stochastic latent variable for masked image prediction: a learned posterior infers z from context and target during training, while a learned context-conditioned prior approximates the distribution of plausible completions at inference, with no decoders and no reconstruction. In its deterministic limit the objective reduces to the I-JEPA loss in functional form, generalising that objective rather than strictly containing the I-JEPA model. Our central finding is that stochasticity adds capabilities, error-correlated predictive uncertainty and diverse completions, at no cost to representation quality: under matched architectures and schedules S-JEPA improves over the deterministic I-JEPA baseline and exceeds LeJEPA at matched architecture, while additionally producing spatially structured uncertainty maps and semantically diverse completions, neither of which the deterministic predictor produces.
Read the full abstract

Spectral Theory

The enigma of locally Toeplitz matrices. I. Discrete Schrödinger operators with complex potential

Freezing the coefficients of a banded locally Toeplitz matrix gives a family of Toeplitz eigenvalue measures. Their average can fill a region of the complex plane even when the eigenvalues of the original matrices appear to follow curves. We conjecture that, when the above average has a sufficiently regular planar density, the limiting eigenvalue measure is its mother body. We prove it for a class of finite Dirichlet discrete Schrödinger operators with complex potential $V$ and a common real coefficient $\alpha>0$ on the two adjacent diagonals. When $(\Im V)'>0$ and $V$ admits a suitable holomorphic extension, the limiting measure is supported on finitely many analytic arcs. Its logarithmic potential falls below that of the averaged measure by twice the least of four explicit actions. We obtain these actions from a four-term asymptotic expansion of the Dirichlet determinant. When $\Im V$ is not monotone, an exact transfer recursion gives a conditional extension. For a circular potential arc we verify the geometric switching condition, leaving non-resonance as the remaining hypothesis.
Read the full abstract

Differential Geometry

Higher Holonomy for Higher Lie Algebroid Connections on Graded Vector Bundles

We prove the existence of a Riemann–Hilbert $A_\infty$-functor from the curved dg-category of $\mathscr{A}$-connections on graded vector bundles, where $\mathscr{A}$ is a higher Lie algebroid, to the curved dg-category of $\Pi_\infty(\mathscr{A})$-connections on graded vector bundles, where $\Pi_\infty(\mathscr{A})$ is the homotopy groupoid of $\mathscr{A}$. When restricted to the sub dg-categories of flat objects, it induces an $A_\infty$-functor from the dg-category of representations up to homotopy of $\mathscr{A}$ to the dg-category of representations up to homotopy of $\Pi_\infty(\mathscr{A})$, which recovers the one from arXiv:1011.4693 and arXiv:0908.2843 when $\mathscr{A}$ is a Lie algebroid. The key ingredient for defining such functors is a non-abelian Stokes's theorem for endomorphism-valued differential forms. Furthermore, for a graded vector bundle $E$, we introduce the notion of its higher general linear groupoid in such a way that $G$-connections on $E$, for a higher Lie groupoid $G$, become smooth functors from $G$ to the higher general linear groupoid of $E$. We combine this with the Riemann–Hilbert functor to define higher parallel transport functors for $\mathscr{A}$-connections. Finally, we use these functors to define the higher holonomy groups of $\mathscr{A}$-connections and endow them with the structure of augmented simplicial Lie groups.
Read the full abstract

High Energy Physics (Phenomenology)

Spin Correlations in $t\bar{t}$ Production and Decay at the LHC in Perturbative QCD

Among recent developments in top-quark physics at the LHC, the production of top-quark pairs near threshold has attracted considerable attention, partly due to tensions between data and theoretical predictions. We show that reliable predictions for threshold-sensitive observables, such as spin correlations and the differential distribution in the $t\bar{t}$ invariant mass, can be obtained in perturbation theory by retaining the first few powers of $\alpha_s/v$, where $v$ is the relative velocity of the top quarks. Applying this approach to reweight nominal Monte Carlo predictions for spin-sensitive observables, we find good agreement with data. We also discuss how threshold effects, including bound-state contributions, can be incorporated on an event-by-event basis into fully differential Monte Carlo event generators.
Read the full abstract

2611.09299 · 6 Nov · 8 pages, 2 figures. With minor differences, it matches the submitted proceedings for talks given at The 33rd International Workshop on Deep Inelastic Scattering and Related Subjects (DIS2026), 4 - 8 May 2026, Bologna, Italy, and The XII edition of QCD@Work, International Workshop on QCD, theory and experiment (QCD@Work 2026), 16 - 19 June 2026, Trani, Italy

Artificial Intelligence

Towards an Autonomous AI Steward for Large-Scale Physics Experiments: Integrating Causal Diagnostics, RAG, and Conformal Prediction

Long-term particle physics experiments face critical operational challenges: monitoring non-stationary machine data streams and preserving tacit expert knowledge. We present the "Helmholtz Co-Pilot," a hybrid Multi-Agent System (MAS) acting as an autonomous experiment steward. The system can be triggered interactively by human operators or autonomously by an asynchronous causal diagnostics engine that runs independently of the MAS reasoning core. Continuously monitoring live hardware streams such as data acquisition and slow control, this engine performs dynamic causal discovery to trace cascading alarms back to their root causes and applies adaptive model selection to keep forecasts reliable across shifting machine states; upon detecting an anomaly, it hands a structured diagnostic report to the MAS. The reasoning layer itself relies on a high-throughput C++ core and a Python edge architecture orchestrating a directed acyclic graph (DAG) of specialized ReAct workers. It integrates a Retrieval-Augmented Generation (RAG) backbone for institutional memory and leverages conformal prediction to bound its evidence-based recommendations with mathematical coverage guarantees. This extended abstract outlines the architecture, the integration of software-enforced Human-in-the-Loop (HitL) guardrails, and ongoing shadow-testing within the KATRIN experiment.
Read the full abstract

All authors Oliver Noll, Nicholas Tan Jerome, Alexander Lehnen, Frank Simon, Frank Maas

High Energy Physics (Phenomenology)

Collins-Soper kernel at large $b$ from Chiral Perturbation Theory

We construct effective operators for transverse-momentum-dependent (TMD) distributions in Chiral Perturbation Theory. The operators constructed from pion and nucleon fields possess the main properties of the QCD operators, including multiplicatively renormalizable rapidity divergences. Extracting the leading rapidity divergence, we find the expression for the Collins-Soper kernel at large $b$ and at the low-energy normalization point, which is model-independent up to an additive constant. In this regime, the Collins-Soper kernel has a nearly constant behavior. We also find that at large $b$ it depends on the isospin channel and on the hadron type, although this dependence is numerically very small.
Read the full abstract

Computation and Language

Multilingual Artificial Datasets for AI-Driven Cardiology: the CliTe and ParaCliTe corpora

Background and Aims. Comparable, multilingual, validated cardiology corpora are critical to advance trustworthy natural language processing (NLP) models for cardiology but such resources are scarce due to privacy concerns. This study aimed to cover this gap by presenting suitable artificial corpus creation guidelines and by releasing 3 multilanguage corpora. Methods. Following purpose-specific publicly-available corpus construction guidelines, cardiologists from 7 European hospitals generated artificial clinical records in Czech, Dutch, English, Italian, Romanian, Spanish and Swedish, ensuring clinical diversity and sex balance, resulting in the CliTe corpus. A CliTe subset underwent a multi-stage manual translation into the other 6 languages, resulting in the parallel ParaCliTe corpus, while the remaining reports were automatically translated, constituting ParaCliTe Silver. Results. CliTe is a collection of 699 artificial cardiology reports, mostly outpatient notes (340 documents), discharge reports (263), and inpatient notes (89), describing fictional patients with ischaemic heart disease (199 patients), heart failure (188), and valvular disease (79), among others. ParaCliTe contains reports from 94 fictitious patients, each available in seven languages with sentence-level alignment, resulting in 658 documents. The corpus mainly comprises discharge reports (322 documents), and outpatient (196) and inpatient (140) notes, describing patients with heart failure (399 documents), ischaemic heart disease (287), and valvular disease (238). Empirical results illustrate the utility of ParaCliTe and highlight the limitations of state-of-the-art NLP models adapted to the cardiology domain. Conclusions. The creation of multilingual, equitable and realistic corpora of artificial patients capable of strengthening NLP performance for cardiology is feasible when supported by detailed guidelines.
Read the full abstract

All authors Lorenzo Pratesi, Eulalia Farre-Maduell, Machteld Boonstra, Michele Miranda, Jonas Mack, Eduard Rodenas-Alesina, Jonathan Kossoff, Corinne Allaart, Sébastien Bratières, Manuela-Daniela Danu, Isabella Fumarulo, Milou Gamage, Antonio Iaconelli, Pavel Leinveber, Felix Lindberg, Pavel Macháč, Vladimíra Macíková, Gianluigi Savarese, Raffaele Scorza, Bram van Es, Emma Weiss, Andrea Guala, Martin Krallinger, the DataTools4Heart NLP group

Neural and Evolutionary Computing

Bridging Spiking and Continuous Dynamics in Spiking Neural Networks: A Muscle-Inspired Model

Transforming the spike trains produced by spiking neural networks (SNNs) into continuous signals through biologically inspired mechanisms remains an open challenge in generative and regression tasks. In biological systems, a similar transformation occurs when muscles integrate neural spikes to produce smooth, coordinated movements. Inspired by this mechanism, we introduce a muscle-response model, a biologically motivated dynamical layer that converts spike trains into continuous representations. Our approach uses second-order dynamics to model the temporal integration of spiking activity. We investigate how the number of spiking neurons and temporal resolution affect the quality of the generated signals in a text-to-speech task, showing that increasing representational capacity has a greater impact than increasing temporal resolution under the evaluated configurations. Furthermore, an analysis of the learned dynamics reveals that, without explicit constraints on the response type, most neurons converge toward non-oscillatory behavior. These results demonstrate the potential of biologically inspired dynamical decoders for bridging spiking neural computation and continuous signal generation.
Read the full abstract

Information Retrieval

Orthogonal Norm Growth in Two-Tower Retrieval: A Popularity Mechanism for Unified Discovery

Unified search and recommendation systems may update the same item vectors from both interaction streams. We study how repeated updates can increase an item vector's norm and affect retrieval under an unnormalized dot product. A cosine-only loss makes the output gradient perpendicular to the current item vector - tangent to its constant-norm sphere. Training changes encoder parameters, however, so the realized item-vector movement need not remain tangent. We derive sufficient conditions under which it does; every nonzero update then increases squared norm by its squared length. When the tasks share an item encoder, sequential updates from both can accumulate in the same norm, which raw-inner-product serving can expose for either task. This conditional result is not a theorem about joint SGD. Controlled ablations test these conditions. On MovieLens 32M and Amazon Books, accumulated tangent movement accompanies frequency-norm association and additional head-item exposure under raw-dot rather than normalized-item ranking. In a nonlinear tower combining a per-item ID vector with a shared content encoding, active item movement remains near-tangent. We separate its norm change into the ID-update contribution and other output changes, then use a matched training intervention that removes immediate norm growth caused by the ID-row update. The intervention reduces norm-related exposure but also recommendation utility, so it is diagnostic rather than a proposed mitigation. Our logs are recommendation-only: the formulation supplies a shared-representation audit, while prevalence in joint traffic remains open.
Read the full abstract

Robotics

MDF++: Visuo-Tactile World Action Model with End-to-end Representation Learning

Recently, world-action models (WAMs) have emerged as a powerful framework for robot manipulation. By jointly modeling action generation and future observation prediction, WAMs can act and anticipate the outcomes of the actions. However, existing WAMs typically rely on frozen pretrained encoders trained with reconstruction, which may not produce representations well suited for robot control. In this paper, we tackle the problem of learning physicallygrounded, control-relevant representations for WAMs. We propose MDF++, a visuo-tactile world-action model that learns latent representations of multisensory observations end-to-end with a diffusion backbone. While naive joint training can lead to feature collapse, MDF++ introduces a multi-task learning strategy that leverages cross-modal correlations to learn task-relevant representations. We evaluate MDF++ on a suite of contact-rich manipulation tasks, including realworld bolt screwing and nut mating, and show that it outperforms WAMs built on frozen pretrained representations by up to 50%. Additional videos are available on our website: jointmdf.github.io.
Read the full abstract

All authors Zixuan Huang, Jinhao Yu, Zachary Charlick, Dmitry Berenson

Numerical Analysis

High-order FEEC discretization on 2D and 3D domains with a disk-like coordinate singularity

Tensor-product splines mapped from a square to a disk-like domain form a discrete de Rham complex, but at the coordinate singularity the pushed-forward $0$-, $1$- and $2$-forms are discontinuous, unbounded and not square integrable. Existing polar constructions impose either high-order regularity on the $0$-forms only, or first-order regularity on the whole complex. For general disk-like mappings composed with a smooth diffeomorphism, we construct polar subspaces of all three tensor-product spaces that are conforming, form a cochain complex with the cohomology of the disk, and are $\mathcal{C}^{\mathfrak{r}-m}$ at the center for $m$-forms, for any $\mathfrak{r}$ up to the spline degree, under a compatibility condition between the mapping and the angular basis. The construction enriches the $2$-form space with projected polynomials and lifts them to curl-like $1$-forms by Poincaré-type homotopy operators, which reduce to cumulative sums of radial B-spline coefficients, so that the homotopy identities hold exactly. The same operators yield sparse commuting projections that preserve total flux and circulation, and boundary-compatible homotopies that solve the curl–curl problem with a single mass-matrix inversion. In 3D, the tensor product with a periodic 1D complex carries all properties over to stellarator-like solid tori. Numerical experiments confirm optimal projection accuracy and the predicted regularity, show that the regularity conditions suppress the stiff central modes of the curl–curl operator, allowing more than ten times larger time steps than first-order polar spaces, and that compatible high-order regularity is required for convergent guiding-center orbits near the axis of a discrete vacuum field.
Read the full abstract

Artificial Intelligence

AutoRefine: Auto-Calibrating Code Judges from Workflow Outcomes

LLM judges increasingly serve as quality gates for code changes that have no test oracle, but a judge prompt must be calibrated for its task and recalibrated as the system it judges evolves, while hand labels for the calibration are resource intensive and go stale. Development workflows emit outcome labels for free (tests pass or fail, changes land or are abandoned), but these labels are noisy. We present AutoRefine, a loop that rewrites a deployable judge's prompt to reproduce outcome labels, with an informed teacher that sees the judge's input plus the outcome, explains each label it keeps, and abstains when the outcome cannot be explained from what the judge sees. On a held-out dataset with an execution oracle, calibration raised chance-corrected agreement (Scott's $\pi$) from 0.36 to 0.60. On a deployed agent's outcome labels, with 25% flipped at random to introduce noise, the teacher set aside flipped fail labels at more than twice the rate of genuine ones, cutting overall label noise from 25% to 20%. Recalibrating that agent's judge using AutoRefine raised its agreement with merge/abandon outcomes on 17 independently held-out diffs from the current workflow ($\pi$ 0.16 to 0.51); on 16 diffs from older workflows, the previous judge stayed ahead, illustrating the need for recalibration as the agent evolves. By abstaining, the teacher not only removes noisy labels but also flags changes whose outcome the judge's input cannot explain, telling a team when a judge needs better inputs rather than a better prompt.
Read the full abstract

All authors Shervin Shahidi, James Poon, Hang Xie, Kenneth O. Xu, Robin Håkanson, Vikram Nair, Jens Pillgram-Larsen, Nachiappan Nagappan

Machine Learning

Safety-Focused Reinforcement Learning Using Risk-Sensitive Coupled Control

This paper investigates the coupled control problem within safety-focused reinforcement learning, addressing the essential challenge of achieving optimal performance while ensuring safety in systems operating within complex environments. We define safety as the requirement to prevent the system from reaching undesirable or hazardous states. We propose a coupled control approach that prioritizes safety by employing effective methods to prevent the system from entering unsafe states while optimizing the performance metric, defined as the average cost. We propose a novel problem transformation that reformulates it into a risk-sensitive problem combined with an average reward criterion. We present a two-timescale Q-learning algorithm where in the faster timescale, the algorithm solves a risk-sensitive control problem to ensure safety, while on a slower timescale, the algorithm focuses on average cost control. We provide theoretical guarantees for the convergence of our algorithms, establishing their significance in ensuring safe, optimal control in dynamic environments. We validated the performance of our approach using GridWorld and CarGoal task in Safety Gymnasium.
Read the full abstract

Analysis of PDEs

Single-scale regularity from Dirichlet energy for the five-dimensional stationary Navier–Stokes equations

We establish an interior $\varepsilon$-regularity criterion for suitable weak solutions of the unforced stationary Navier–Stokes equations in an open set $\Omega\subset\mathbb R^5$. There exists a universal constant $\varepsilon_\ast>0$ such that $x_0$ is regular if, for some ball $B_R(x_0)\subset\Omega$ with $R>0$, \[ \frac1R\int_{B_R(x_0)}|\nabla u|^2\,dx \leq \varepsilon_\ast^2. \] The main difficulty is that the Dirichlet energy controls the velocity oscillation while its mean may be arbitrarily large. Subtracting the mean introduces a large constant drift into the perturbation equation. Moreover, smallness in $H^1$ alone is insufficient to close a direct nonlinear elliptic estimate in five dimensions. We overcome these difficulties through uniform Oseen estimates and a relative-energy comparison yielding quadratic improvement of the perturbation. As a consequence, the singular set has zero one-dimensional upper Minkowski content on every compact subset of $\Omega$, and hence local upper Minkowski dimension at most one.
Read the full abstract

Applied Physics

Absolute Calibration of a Broadband Transmission Grating Spectrometer from 5 to 400 nm with Uncertainty Estimates

We present the absolute calibration of a transmission grating spectrometer (TGS) with uncertainty estimates in the 5–400 nm range using synchrotron radiation at the BEAR beamline at Elettra light source in Trieste, Italy. Firstly, the spectral responsivity of a Si photodiode in the spectrometer was absolutely calibrated with respect to the absolutely-calibrated beamline photodiode. Subsequently, this photodiode was used to calibrate the transmittance of Zr and Al thin-film filters, along with bulk MgF$_2$ and UV-grade fused silica windows at their operating positions in the spectrometer to minimize mounting-induced uncertainties. Next, the quantum efficiency of the spectrometer's CCD camera was calibrated over the 5–400 nm range. Lastly, three gratings with line densities of 10000, 3000, and 1000 lines/mm were calibrated for spectroscopy in the ranges extending from soft X-ray to extreme-, vacuum- and deep-ultraviolet (DUV) wavelengths, respectively. The combination of various filters and gratings enables absolutely calibrated spectroscopy with high versatility across a broad wavelength range from soft X-rays to DUV.
Read the full abstract

All authors E. Chebykin, A.E.H. Gaballah, R. van der Horst, A. Giglia, N. Mahne, M. van de Kerkhof, M.D. Ackermann, M. Bayraktar

Statistical Mechanics

Thermalization of coarse-grained variables in the Toda lattice: a numerical study

We present here a numerical study of the Toda lattice Hamiltonian dynamics showing how, despite model integrability, it is possible to find many macroscopic observables which display standard equilibrium behavior. In particular, we study numerically the energy time correlation of certain coarse-grained variables, showing that it decays with either exponential or stretched exponential behaviour, thus indicating good thermalization properties. The spirit of our observation is complementary to that of standard Fermi-Pasta-Ulam-Tsingou-like numerical experiments, in which the breaking of integrability is regarded as the key ingredient for relaxation to thermal equilibrium. The present communication corroborates, on the contrary, the view that integrability is not an obstacle to thermalization, provided that a sufficiently weak definition of the latter is accepted. In the Toda lattice we have introduced several kinds of coarse-grained variables, constructed with the purpose to maintain a thermodynamic relevance also for large system sizes, which all show good thermalization properties.
Read the full abstract

Statistical Finance

The End-of-Day Reversal After 2019: A Pre-Registered Test on S&P 500 Stocks Excluding Heavily Traded Names

Baltussen, Da and Soebhag (2025) document an end-of-day reversal in U.S. stocks through December 2019: intraday losers outperform intraday winners over the last half hour of trading. We test it on later data, on point-in-time S&P 500 members excluding about 80 heavily traded stocks we had examined in earlier, unpublished work. Under a private pre-registration, the 2016-2019 data had to reproduce the published sign before any statistic linking the sort to the outcome was computed on later primary-sample data. Measured to the last trade before the closing auction, the dollar-volume-weighted reversal is 3.13 bps per day (t = 7.5) in 2016-2019 and 3.05 bps (t = 3.0) in 2023-2026, with a one-sided 95% bound on the decline of 61%. The result depends on the window: the reversal is indistinguishable from zero in 2020-2022, 2023 carries much of the later estimate, and from 2020 onward it is 1.69 bps (t = 2.3), with a bound of 91%. A proxy sample, added after the run, gives 2.81 and 3.86 bps. On the full index the reversal falls from 3.40 to 0.74 bps (t of the change -1.83, short of the registered bar of -3); descriptive comparisons concentrate this weakness in the very largest stocks. No pre-registered mechanism test passes.
Read the full abstract

2611.09286 · 6 Nov · 15 pages, 1 figure, 9 tables. Replication archive (code, point-in-time membership, leveraged-ETF launch table, pre-registration documents): https://doi.org/10.5281/zenodo.24003093. The paper makes no claim net of costs and does not propose a trading strategy

Mesoscale and Nanoscale Physics

Momentum-space engineering in van der Waals heterostructures

Twenty years of graphene research have established Dirac fermions as a versatile platform for exploring quantum topological phenomena. Yet control over graphene's low-energy spectrum mostly follows the route of miniband engineering using real-space superlattices. Here we introduce a complementary approach based on momentum-space 'Umklapp' band engineering. This is realised by matching reciprocal space positions of band edges of two adjacent highly incommensurate 2D crystals to coincide upon a boost from their Bragg vectors, creating resonant interlayer hybridisation that qualitatively reconstructs their spectra. To demonstrate this approach experimentally, we use monolayer graphene on MoS$_2$, twisted such that MoS$_2$ Q-valley states are brought by Umklapp into energy-momentum resonance with graphene's Dirac cones. The ensuing hybridization results in the formation of a clear additional Dirac-like spectral feature, identified by a distinct Landau-level fan and Shubnikov-de Haas oscillations characteristic to Berry phase $\pi$ electrons. These new topological spectral features create a spin-valley locked Ising system which is robust against small twist-angle variations that can be further tuned by an external displacement field. More broadly, Umklapp band engineering offers a practical strategy for tailoring the electronic structure in van der Waals heterostructures, expanding the design space of quantum materials beyond moiré superlattices and weak proximity effects.
Read the full abstract

All authors Xiao Li, Astrid Weston, James G. McHugh, Michael D. Thompson, Matthew Watson, Yunze Gao, Nicholas Clark, Laxman Nagi Reddy, Samuel Magorrian, Wendong Wang, Zhiyuan Wang, Sofia Konyzheva, Francisco Selles, Amy Carl, Katarzyna J. Salamon, Alex Summerfield, Nicholas Hine, Andre K. Geim, Neil R. Wilson, Jonathan R. Prance, Andrey V. Kretinin, Vladimir I. Falko, Roman Gorbachev

Algebraic Geometry

Dominant root sequences of arbitrary degree

We show that for root-subgroup multiplication maps, dominance does not imply birationality. This gives a negative answer to a question of Fang, Fourier and Littelmann. More precisely, for every integer $r\geq 2$, we construct a sequence of positive roots in type $A_{3r-1}$; the associated multiplication map is dominant of generic degree $r$, hence not birational. In our construction, a key observation is that Lusztig's flag incidence projection for a triangle quiver is, up to birational equivalence and identity factors, simultaneously a root-subgroup multiplication map and a multiplicative partial Grothendieck–Springer map. Moreover, we extend these examples to types $B$, $C$ and $D$ in every generic degree $r\geq 2$, and to types $E_6$, $E_7$ and $E_8$ in generic degree $2$.
Read the full abstract

Optimization and Control

Hierarchical Signal Temporal Logic Control of Continuous-Time Systems

For linear continuous-time systems, we solve the signal temporal logic (STL) control problem by translating it from continuous to discrete time such that a solution to the discrete-time problem guarantees a solution to the original continuous-time one. We split control synthesis into three hierarchical tasks. First, we discretize the continuous-time STL formula into a discrete-time one by sampling the temporal behavior it specifies. We accordingly discretize the continuous-time system by constructing a discrete-time model such that the input-state trajectories of the continuous-time system can be characterized as piecewise polynomial interpolations of the input-state trajectories of the discrete-time model. Second, we exploit this characterization to formulate mixed-integer linear constraints on the discrete-time trajectories that allow for planning a trajectory whose piecewise polynomial interpolation satisfies the continuous-time STL formula. In fact, system discretization and trajectory planning are performed in such a manner that any feasible solution to these constraints gives a discrete-time trajectory that can be transformed into a continuous-time one guaranteed to satisfy the original STL formula. Third, we propose a step-by-step algorithm to construct the control input that realizes such continuous-time trajectory.
Read the full abstract

Artificial Intelligence

Ontology-Driven Control of Conversational Agents with a Use-Case for Job Interview Practice

Large Language Models (LLMs) have been widely used as conversational agents in both domain-specific and open-domain situations. In domain-specific context, it is often necessary to adapt those LLMs to obtain conversations with desired characteristics, whether those apply at the level of each response (e.g. for tone or style) or more globally (i.e. at the level of a conversational strategy). Such adaptations are, however, generally realized in an ad hoc way, through specific, non-reusable, and non-transparent developments. In this paper, we propose an ontology-based approach to design conversational strategies that rely on the desired and measured characteristics of textual utterances in the conversation. Through a use case related to job interview practice, we show how this approach can be used to control LLMs to follow the strategy defined in the ontology. To this end, we automatically fine-tune the LLM to improve its compliance with ontologically-defined conversation characteristics. Our results show that training the LLM from formally defined concepts effectively enhances alignment with the requested characteristics. Hence, this work demonstrates how ontology-guided conversation control can be achieved while taking into account multiple, concurrent conversation aspects that are formally and transparently defined.
Read the full abstract

Numerical Analysis

A semi-implicit exactly well-balanced relaxation scheme for the 2d shallow water system

We propose a semi-implicit, exactly well-balanced finite volume scheme for the two-dimensional shallow water equations over a non-flat bottom, aimed at the low-Froude-number regime. The method combines a Suliciu-type relaxation of the pressure with a flux splitting that separates the fast acoustic waves from the slow transport ones: the pressure subsystem is integrated implicitly and the transport subsystem explicitly, so that the time step is no longer restricted by the acoustic stability condition. Written in terms of its Riemann invariants, the pressure subsystem decouples into independent one-dimensional problems along the grid lines, each of them governed by a banded linear system whose bandwidth does not depend on the mesh. The implicit step therefore requires no non-linear solver and costs O(NxNy) operations, regardless of the acoustic Courant number. A two-dimensional exactly well-balanced reconstruction, whose limiter weights are frozen within each time step, preserves this linear structure for the second-order scheme. We prove that both the first- and the second-order schemes preserve water-at-rest equilibria exactly, and we assess them on several two-dimensional benchmarks, where they resolve fine-scale velocity structures on coarse meshes and run up to 8.5 times faster than standard explicit Riemann solvers.
Read the full abstract

All authors José Ignacio Ramírez-Fuentes, Cipriano Escalante, Tomás Morales de Luna, María Luz Muñoz-Ruiz

Artificial Intelligence

An Empirical Study of Issues in GPU Kernel Libraries for LLM Systems

Behind the expanding use of large language models (LLMs) across application domains lies a growing dependence on specialized GPU kernel libraries. These libraries underpin the correctness and efficiency of training and inference across diverse GPU platforms, implementing attention, GEMM, quantization, and low-precision operators while managing tensor shapes, memory layouts, code generation, and hardware-specific dispatch. Problems at this layer can silently corrupt outputs, hang GPUs, prevent compilation, introduce deployment incompatibilities, or degrade performance, thereby reducing developer productivity and wasting expensive GPU resources. We present the first comprehensive empirical study of issues in GPU kernel libraries for LLM systems. We manually analyzed 300 high-quality GitHub issue reports, their discussions, and available fixes from seven representative libraries: FlashInfer, FlashAttention, Triton, CUTLASS, DeepGEMM, xFormers, and bitsandbytes. The analysis characterizes symptoms, root causes, fixes, and testing and debugging practices. Our main findings are: (1) 60.33% of the sampled issues originate in the Compiler and Build System or Runtime Environment dimensions, so validation must cover the deployment and execution stack, not just kernel code. (2) Kernel specialization imposes implicit correctness constraints on tensor shapes, metadata layouts, memory aliasing, and hardware resources. Violations can produce incorrect outputs or launch failures even for supported operations, highlighting the need for explicit contract validation and boundary-case testing. (3) Current testing and debugging practices are often insufficient for identifying low-precision numerical problems and hardware-specific execution failures. We outline research opportunities and tooling directions to support future LLM GPU kernel development.
Read the full abstract

All authors Yanjie Gao, Jiayi Bi, Haoxiang Lin, Ling Feng

Sound

AnomSep: Bridging Optimization-Based Separation and Deep Neural Refinement for Anomaly Detection with Listenable Explanations

We propose AnomSep, which combines optimiza-tion-based separation with deep neural refinement to achieve accurate anomaly detection with listenable explanations. Anomaly detection requires no predefined anomalous categories, and understanding the basis for each detection helps security personnel assess the validity of each detection. End-to-end anomaly detectors can achieve high detection accuracy but output only scalar scores, and existing anomalous sound separation methods provide listenable explanations with limited detection accuracy. AnomSep first obtains preliminary normal-sound and anomalous-sound estimates through optimization-based separation using a dictionary learned from normal sounds in the target environment. Deep neural refinement then improves the preliminary estimates in a pretrained audio latent space. The separated anomalous sound serves as the listenable explanation, and its energy determines the anomaly score. Experiments on synthetic mixtures and real mixtures show improved separation quality over baseline separators. AnomSep achieves detection performance competitive with the end-to-end baseline and provides listenable explanations directly linked to each detection decision.
Read the full abstract

All authors Koki Shoda, Jun Younes Louhi Kasahara, Qi An, Atsushi Yamashita

Machine Learning

Diversity-Aware Client Selection for Communication-Efficient Federated Learning

Federated learning (FL) enables collaborative model training across distributed clients without sharing raw data, making it attractive for privacy-sensitive and bandwidth-constrained applications. A practical challenge in deploying FL is reducing communication overhead while maintaining model performance under heterogeneous, non-IID data distributions. Selecting which subset of clients participates in each training round offers a principled way to address both challenges simultaneously, yet its impact on data diversity and representativeness remains underexplored. In this paper, we investigate three client selection strategies, Power-of-Choice, Fisher information, and Centered Kernel Alignment, evaluating them not only by accuracy and convergence, but through the lens of participation diversity using the Gini coefficient and KL divergence. Under non-IID conditions, restricting participation to 15% of clients per round achieves significant communication savings, and when selection promotes diverse and balanced contributions, it also improves generalization over biased alternatives. This work provides a systematic and reproducible analysis demonstrating that diversity-focused client selection renders communication efficiency and model quality complementary rather than competing objectives.
Read the full abstract

All authors Max Halleryd, Sima Sinaei, Viktor Valadi, Mina Alibeigi, Henrik Abrahamsson, Salman Toor

Cryptography and Security

Federated Forensics: Jurisdictional constraints as a design driver for multi-agency federated learning

Digital forensic investigations increasingly cross borders, and law enforcement agencies need to build analytical models together without centralizing their data. Federated Learning (FL) keeps data local, yet neither FL nor its frameworks enforce the jurisdictional, governance, and accountability rules that multi-agency forensic work requires. This paper takes a design-first approach to that gap. First, we derive seven requirements that a federated learning system must meet in a forensic setting. Second, we propose Federated Forensics, a three-layer reference architecture whose middle layer, unlike in existing hierarchical FL designs, sits at a jurisdictional border and can use a model update locally while blocking it from crossing that border. Third, we build a prototype of the architecture and test it on a malware-family classification task. In a prototype evaluation, the policy gate blocked every restricted update at the border with zero violations, and restricted-class accuracy rose from 0% under today's compliant alternative to 93.0%, only 4.4 points below ignoring the restriction.
Read the full abstract

All authors Aruna Withanage, Jessica Herrmann, Hudan Studiawan, Frank Breitinger

Systems and Control

DC-Link Voltage Regulation in a Single-Isolation-Stage Multioutput Solid-State Transformer for Wide-Range DC Outputs

Single-isolation-stage solid-state transformers with balancing circuits eliminate an additional downstream dc–dc conversion stage. However, the remaining isolated converters must directly accommodate a wide output-voltage range. This paper investigates variable dc-link voltage operation of a single-isolation-stage multioutput SST to ease this requirement. Coordinated cascaded H-bridge modulation and embedded four-switch balancing circuits support intentionally unequal submodule dc-link voltages, allowing the isolated-converter input voltages to be adjusted according to the output requirements. The achievable voltage range is determined by relating the required balancing power to the BC transfer capability within the available CHB switching windows, subject to modulation and peak and rms current limits. The analysis quantifies the range extension provided by BC operation beyond modulation alone. For a single-phase-shift dual-active-bridge converter, input-voltage following within this range reduces the calculated peak current and extends the ideal zero-voltage-switching region compared with fixed input. Experiments on a six-submodule SST prototype demonstrate unequal dc-link voltage operation and inter-submodule power exchange. Experiments result on DAB demonstrate reduced current stress and improved efficiency with voltage following.
Read the full abstract

All authors Yi Hao, Levy Ferreira Costa, Krzysztof Puczko, Liangcai Shu, Dongsheng Yang