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

Saturday 12 September

Algebraic Geometry

Log Canonical Counterexamples to the Chen–Jiang Decomposition

We construct a projective log-smooth surface pair $(X,\Delta)$ and a morphism $f:X\to E$ to an elliptic curve such that $f_*\OO_X(K_X+\Delta)$ is Atiyah's indecomposable rank-two bundle $F_2$ of degree zero. This gives a negative answer to Question 3.4 of Jiang \cite{Jiang}. We also construct a log canonical pair $(X,\Delta)$ and a morphism $f$ to an abelian surface such that $f_*\OO_X(m(K_X+\Delta))$ does not admit a Chen–Jiang decomposition for any integer $m\ge2$. Products yield counterexamples in every higher direct image degree and over abelian bases of arbitrary positive dimension.
Read the full abstract

Information Theory

Worse Routes, Fresher Information: Preemption Redefines Which Route is Best

This paper studies joint sampling, route activation, and preemption control over $N$ heterogeneous transmission routes under a long-run objective combining Age of Information (AoI) with sampling and transmission costs. We formulate the problem as an average-cost impulse-control model, derive a vector-form integral average-cost optimality equation over a hybrid state space, and establish structural reductions on the state. We consider a two-route case and compute optimized policies using discretized policy iteration. Surprisingly, the results show that preemption can reverse routing preferences: the optimized policy may exclusively use a route with both a larger mean delay and a larger delay variance while avoiding the seemingly superior route. Relative to baselines that do not allow preemption, the proposed policy reduces average AoI by up to $96\%$.
Read the full abstract

Medical Physics

Fail-closed conformal multiresolution error control for conservative three-dimensional dose remapping in synthetic phantoms

Conservative dose remapping can be formulated by transporting mass and deposited energy across nonmatching grids, but fixed sampling uses the same number of points per active source cell regardless of case difficulty. We developed conformal multiresolution error control (CoMERC), a fail-closed multilevel quasi-Monte Carlo (QMC) controller for reference-relative target-cell allocation error in a fixed piecewise-constant source model with known deformation. Two randomized nested replicates were evaluated at a probe level of 4 and candidate levels of 8 and 16. Four endpoints were controlled jointly: reference-mass-weighted global, high-gradient, and density-gradient-threshold root-mean-square error, plus eligible-cell maximum absolute error. A split-conformal multiplier was calibrated from 240 synthetic cases, frozen, and evaluated on 400 independent cases from the same generator and 120 cases from a prespecified shifted generator. In the primary set, joint coverage was 394/400 (98.50%; one-sided 95% lower limit, 97.06%), 388/400 cases (97.00%; lower limit, 95.18%) received an output, and 0/388 released outputs exceeded any tolerance (one-sided 95% upper limit, 0.769%). Final actions were level 8 for 190 cases, level 16 for 198, and ABSTAIN for 12. The mean of the case-normalized, policy-implied production-sample ratios was 0.5844, with each case normalized to always running both randomized replicates through level 16; the prespecified 95th-percentile bootstrap upper limit was 0.6194. All seven primary gates passed. In the shifted-generator set, 113/120 cases were released and one released output exceeded a tolerance. CoMERC provided marginal finite-sample error control and lower modeled production-sample work in the locked synthetic population, but not patient-level, registration-level, conditional-on-release, or clinical safety validation.
Read the full abstract

Functional Analysis

Local-to-Global Isometries in Standard-Norm Generalized Gyrovector Spaces: Thompson and Hilbert Geometries of JB-Cones

We introduce standard-norm generalized gyrovector spaces: generalized gyrovector spaces whose norm-value line has the ordinary real operations and whose fixed scalar maps are continuous. We prove that every bijective gyrometric isometry between nonempty connected open subsets of two such spaces extends uniquely to a global bijective gyrometric isometry. The proof combines gyromidpoints, point reflections, continuation of local isometry germs, and monodromy; joint continuity of scalar multiplication is derived, rather than assumed. The class includes all nonzero real normed spaces and standardized Möbius, Einstein, and proper-velocity models, and is stable under finite and countable $\ell^p$-products. We also obtain bounded mapping-space examples and, under natural continuity assumptions, standard-norm function-space GGVs $C(K,G)$ for compact $K$. These yield non-Hilbertian and hybrid examples. For every unital JB-algebra, the positive invertible cone carries a one-parameter family of standard-norm GGV structures whose gyrometrics are conjugate to Thompson's metric by power maps. The projective positive cone has a parallel family, based on the variation norm modulo the unit, conjugate to Hilbert's projective metric. The projective structures are contractible. Together with global JB-cone isometry classifications, the abstract theorem yields local-to-global extension and classification results for Thompson and Hilbert geometries. Full-rank density matrices and normalized state spaces of finite-dimensional Euclidean Jordan algebras, including the Albert algebra, give concrete projective examples.
Read the full abstract

Materials Science

Physics-aware global Rietveld refinement for high-energy X-ray diffraction microscopy with application to reconstructing intragranular orientation and strain fields

High-energy X-ray diffraction microscopy (HEDM) has emerged as a critical technique for studying the microstructure and, increasingly, strain fields in solids. However, current algorithmic or experimental methods to obtain intragranular fields are time-intensive, provide limited spatial resolution, or yield stress and strain fields that do not satisfy the universal laws of deformation (compatibility and equilibrium). In the context of standard HEDM, a novel physics-aware approach is presented in which the physics of deformation is included in the forward diffraction simulation to ensure that the reconstructed fields are physically meaningful. The entire simulated and experimental diffractograms are compared with a differentiable optimal transport–type objective, and a Rietveld refinement is carried out globally on the internal fields and grain topology using gradient-based optimization. The method is developed, verified with synthetic data, and demonstrated experimentally using near-field HEDM data from aluminum oxynitride (a brittle ceramic), with a reference implementation released as PARA-X. The reconstructions show remarkable improvement over existing methods (improved completeness and loss), and the high-fidelity, high-resolution recovery paves the way for using HEDM to study fine-scale deformation mechanics over large polycrystalline volumes.
Read the full abstract

All authors Carter K. Cocke, Eitan Camacho, Sara F. Gorske, Katherine T. Faber, Kaushik Bhattacharya

Analysis of PDEs

1D Convection-Diffusion in Porous Media: A Unified Solution from Bounded Domain to Half-Space

We present an analytical solution for 1D atmospheric pollutant dispersion in a bounded domain. Using coordinate transformation and spectral decomposition, we obtain a Fourier series solution for time-dependent wind velocity. For constant wind, Laplace inversion yields dual representations via residue calculus and Poisson summation. The solution extends to semi-infinite domains, recovering classical error-function profiles.
Read the full abstract

Computer Vision

ChatGPT Images 2.5 on Forgery Tasks: Testing Advertised Improvements Against Known Answers

OpenAI released ChatGPT Images 2.5 on 8 September 2026, advertising more precise local edits, better consistency across edits, more faithful reference products and sharper detail. We evaluate these claims on four forgery tasks with answers fixed in advance: receipt-field alteration, repeated editing, product placement and small-print rendering. GPT-Image-2 provides same-week baselines at a cheaper and a more expensive tier. A limited improvement appears in receipt editing. After alignment, OCR detects changes to surrounding text in 31.7% of Flare outputs, against 44.2% for the cheaper baseline. This gain is concentrated on CORD receipts and sensitive to shifts of a pixel or less; the forged value itself is no more often correct. Repeated editing and fine print show no measurable gain. Product codes become more legible mainly because Images 2.5 draws the product larger. Defence outcomes change little: localisation remains weak for both generations. A detector that flags 68.6% of controlled benchmark images flags only 35.9% of images posted online. Advertised improvements therefore transfer unevenly to the tested forgery capabilities, while substantial detection limitations remain.
Read the full abstract

All authors Ankit Raj, Yuxin Zhang, Kidus Zewde, Tommy Duong, Jiaqi Gan, Xingyu Shen, Yuchen Zhou, Huaiyu Guo, Siyu Zhang, Simiao Ren

Chaotic Dynamics

First-order integrability-breaking phase transitions in dynamical systems

We investigate a discontinuous route from integrability to chaos using a confined stochastic random walk and a deterministic stadium-like billiard. In both systems, the stationary diffusive observable exhibits a finite jump at the transition: it vanishes at the unperturbed limit but approaches a finite, geometry-controlled value for arbitrarily small nonzero perturbations, providing the characteristic order-parameter signature of a first-order transition. Despite this discontinuity, the relaxation timescale diverges as the transition is approached, revealing critical slowing down. Both models exhibit normal diffusion with $\beta=1/2$, a perturbation-independent stationary state with $\alpha=0$, and a crossover iteration scaling as $n_x\propto\lambda^{-2}$, yielding $z=-2$, where $\lambda$ denotes the corresponding perturbation parameter. The common exponent set $(\alpha,\beta,z)=(0,1/2,-2)$ originates from the same coarse-grained mechanism: normal diffusion within a finite accessible domain with a diffusion coefficient that vanishes quadratically at the transition. The agreement between stochastic transport and deterministic chaotic scattering provides strong evidence for a common class of discontinuous dynamical transitions and extends the statistical-mechanics description of phase transitions to integrability-breaking dynamics.
Read the full abstract

All authors Anne Ketri P. da Fonseca, Marcelo de Almeida Presotto, Diego F. M. Oliveira, Edson D. Leonel

Computation and Language

The University of Melbourne WMT 2026 CreoleMT Submission: A Domain-Balanced Approach to Low-Resource Pacific Creole Machine Translation

For our submission to the WMT26 Creole Language Translation Shared Task, we focus on machine translation (MT) models for Pacific creoles: Tok Pisin, Bislama, and Solomon Pijin, with particular attention to broad domain performance. After pre-training on a large collection of domain-imbalanced data, we continue fine-tuning on a diverse mix of domain-balanced data. We rely on a number of data collection and preparation techniques, including LLM-assisted respelling and alignment, back-translation, and distillation from Gemini for domains originally not present in training data. Evaluated on Bouquet and a novel test set made of spoken language transcripts, our models beat open model baselines by 3+ chrF++ points in all directions with human-original references. Looking ahead, we plan to develop human-translated test sets for Solomon Pijin and Bislama, and to distil our best models into much smaller ones that retain broad domain coverage.
Read the full abstract

Chaotic Dynamics

First-Order Transition to Chaos with Critical Slowing Down

Can the transition from integrability to chaos be discontinuous? We show that it can, and that the resulting first-order dynamical transition coexists with critical slowing down. Using an analytically tractable confined random walk and a deterministic stadium-like billiard, we find a finite jump of the stationary diffusive observable at the transition while the relaxation time diverges. Both systems display normal diffusion and the same exponent set $(\alpha,\beta,z)=(0,1/2,-2)$. We trace this agreement to a common coarse-grained mechanism: diffusion in a finite accessible domain with a diffusion coefficient that vanishes quadratically with the perturbation. The results identify a discontinuous route from integrability to chaos and provide evidence for a broader universality class of first-order dynamical transitions.
Read the full abstract

All authors Anne Ketri P. da Fonseca, Marcelo de Almeida Presotto, Diego F. M. Oliveira, Edson D. Leonel

Theoretical Economics

Cautious and fair social evaluation under risk

In this paper, we study criteria for social evaluation of resource distributions under risk when individuals have expected utility preferences but their utility functions are interpersonally non-comparable. In particular, we focus on cautious evaluation with respect to ex-post distributions. As a key rationality axiom, we adopt Negative Certainty Independence, introduced by Dillenberger (2010). Using this axiom together with Pareto conditions that are compatible with both the rationality axiom and ex-post equity, we characterize a class of social criteria called Cautious Equally Distributed Equivalent (CEDE) criteria. These criteria evaluate risky situations by the lowest certainty equivalent over a set of social utility functions constructed from individual utility functions.
Read the full abstract

Friday 11 September

Representation Theory

On the Classification of Nilpotent Lie Superalgebras over Fields of Characteristic 2

The salient result of this paper is the complete classification of all nilpotent Lie superalgebras in dimensions up to four over an algebraically closed field of characteristic 2. Our method adapts the Skjelbred-Sund method to show that every n-dimensional nilpotent Lie superalgebra arises as a one-dimensional central extension of an (n-1)-dimensional one, and determines precisely when two such extensions yield isomorphic superalgebras. Carrying out this construction for dimensions up to 4, we obtain a finite, non-redundant list of pairwise non-isomorphic examples.
Read the full abstract

Computation and Language

AI-Driven SaaS Platform Integration: AI Express for Unified Generative AI Services with Pay-As-You-Go Credit Systems and Scalable Microservice Architecture

Generative AI is currently fragmented, with separate subscriptions required for text generation, image synthesis, and code assistance, leading to high costs and workflow disruptions. AI Express is a SaaS platform that consolidates these capabilities into a unified pay-as-you-go service. The platform provides a range of AI tools, including a Core Code Reviewer that evaluates GitHub repositories, a Text-to-Website generator that converts text into fully functional websites, script generation, grammar checking, data analysis, and watermark removal, all accessible through a single interface without requiring technical expertise.
Read the full abstract

Instrumentation and Detectors

The Skin Detector System of the LZ Dark Matter Experiment

This paper presents the design, calibration, and performance of the Skin Detector component of the LUX-ZEPLIN (LZ) dark matter experiment. The Skin Detector consists of approximately 2 tonnes of liquid xenon instrumented with 131 photomultiplier tubes, surrounding the central time-projection chamber and functions as an active veto system for background rejection. Energy calibration with monoenergetic gamma sources established a position-dependent energy scale and determined prompt and delayed veto thresholds. Radon decays in the Skin can be identified with less than 15~mm spatial resolution, revealing stable convective flow patterns with measured velocities ranging from 2-4 mm/s that persisted throughout extended data-taking periods. The achieved Skin veto threshold meets or exceeds the design goal of 100~keV throughout the Skin volume. The prompt veto efficiency for gamma rays emitted following electron capture in $^{127}$Xe is found to be ($88 \pm 2$)%, consistent with the simulated value of ($90.4 \pm 0.6$)%. M-shell EC decays are used to validate the LZ detector sensitivity for low-mass dark matter during the first science run.
Read the full abstract

All authors D.S. Akerib, A.K. Al Musalhi, F. Alder, B.J. Almquist, C.S. Amarasinghe, A. Ames, T.J. Anderson, N. Angelides, H.M. Araújo, J.E. Armstrong, M. Arthurs, A. Baker, S. Balashov, J. Bang, J.W. Bargemann, E.E. Barillier, K. Beattie, A. Beauchene, A. Bhatti, T.P. Biesiadzinski, H.J. Birch, E. Bishop, G.M. Blockinger, C.A.J. Brew, P. Brás, S. Burdin, M.C. Carmona-Benitez, M. Carter, A. Chawla, Y.T. Chin, S. Contreras, M.V. Converse, R. Coronel, A. Cottle, G. Cox, D. Curran, C.E. Dahl, I. Darlington, S. Dave, A. David, J. Delgaudio, S. Dey, L. de Viveiros, L. Di Felice, C. Ding, J.E.Y. Dobson, E. Druszkiewicz, S. Dubey, C.L. Dunbar, S.R. Eriksen, N.M. Fearon, N. Fieldhouse, S. Fiorucci, H. Flaecher, E.D. Fraser, T.M.A. Fruth, P.W. Gaemers, R.J. Gaitskell, A. Geffre, J. Genovesi, C. Ghag, J. Ghamsari, A. Ghosh, S. Ghosh, R. Gibbons, S. Gokhale, J. Green, M.G.D. van der Grinten, J.J. Haiston, C.R. Hall, T. Hall, R.N. Hampp, S.J. Haselschwardt, M.A. Hernandez, S.A. Hertel, G.J. Homenides, M. Horn, D.Q. Huang, D. Hunt, E. Jacquet, R. James, K. Jenkins, A.C. Kaboth, A.C. Kamaha, M.K. Kannichankandy, D. Khaitan, A. Khazov, J. Kim, Y.D. Kim, D. Kodroff, E.V. Korolkova, H. Kraus, S. Kravitz, L. Kreczko, V.A. Kudryavtsev, T.A. Kumar, C. Lawes, E.B. Leon, D.S. Leonard, K.T. Lesko, C. Levy, Z. Li, J. Lin, A. Lindote, W.H. Lippincott, J. Long, M.I. Lopes, W. Lorenzon, C. Lu, S. Luitz, V. Mahajan, P.A. Majewski, A. Manalaysay, R.L. Mannino, R.J. Matheson, C. Maupin, M.E. McCarthy, D.N. McKinsey, J. McLaughlin, J.B. McLaughlin, R. McMonigle, B. Mitra, E. Mizrachi, M.E. Monzani, K. Morå, E. Morrison, B.J. Mount, M. Murdy, A.St.J. Murphy, H.N. Nelson, F. Neves, A. Nguyen, C.L. O'Brien, F.H. O'Shea, I. Olcina, K.C. Oliver-Mallory, J. Orpwood, K.Y. Oyulmaz, K.J. Palladino, N.J. Pannifer, S.J. Patton, B. Penning, G. Pereira, E. Perry, T. Pershing, A. Piepke, S.S. Poudel, Y. Qie, J. Reichenbacher, G.R.C. Rischbieter, E. Ritchey, H.S. Riyat, R. Rosero, N.J. Rowe, T. Rushton, D. Rynders, S. Saltão, D. Santone, A.B.M.R. Sazzad, R.W. Schnee, G. Sehr, B. Shafer, S. Shaw, W. Sherman, K. Shi, T. Shutt, C. Silva, G. Sinev, J. Siniscalco, A.M. Slivar, A.M. Softley-Brown, R. Smith, V.N. Solovov, P. Sorensen, J. Soria, T.J. Sumner, A. Swain, M. Szydagis, D.R. Tiedt, D.R. Tovey, J. Tranter, M. Trask, K. Trengove, M. Tripathi, A. Usón, A.C. Vaitkus, O. Valentino, V. Velan, A. Wang, J.J. Wang, Y. Wang, L. Weeldreyer, T.J. Whitis, K. Wild, M. Williams, J. Winnicki, L. Wolf, F.L.H. Wolfs, S. Woodford, D. Woodward, C.J. Wright, Q. Xia, J. Xu, Y. Xu, M. Yeh, D. Yeum, J. Young, H.D. Zenger, W. Zha, H. Zhang, T. Zhang, Y. Zhou

Statistics Applications

CircaFMM: A Flexible Statistical Framework for Circadian Rhythmicity Detection in Transcriptomics

Detecting circadian rhythmicity in large-scale transcriptomic data is challenging when sampling is irregular, signals are weak, and expression profiles deviate from sinusoidal shapes. The fixed harmonic form of classical Cosinor-based methods may therefore be poorly suited to asymmetric and temporally complex rhythms. We present CircaFMM, a flexible framework for circadian rhythmicity detection based on the single-component Frequency Modulated Möbius (FMM) model. CircaFMM defines rhythmicity as time dependence induced by the Möbius phase parameter, so that the non-rhythmic null corresponds to a time-invariant mean profile. The resulting non-regular test combines time-density weighting, regularized estimation, and bootstrap calibration of a weighted likelihood-ratio-type statistic. Simulations showed controlled type-I error under irregular sampling and improved power over the classical Cosinor F-test for asymmetric waveforms. Applied to the GTEx-13T postmortem transcriptomic dataset, CircaFMM recovered canonical clock genes and circadian pathways while identifying additional rhythmic candidates with lower signal-to-noise-ratio, greater waveform asymmetry, and non-sinusoidal temporal profiles. CircaFMM is available at https://circafmm.shinyapps.io/circafmm_shinyapp/, with source code at https://github.com/FMMGroupVa/CircaFMM.
Read the full abstract

Human-Computer Interaction

Sycophancy Through a Five-Level AI Response Validation Framework

AI sycophancy, the tendency of AI systems to excessively agree with, flatter, or validate users, is an emerging concern in human-AI interaction. It is especially consequential in problem-sharing contexts, where users may seek both information and emotional validation. This paper conceptualizes AI sycophancy as excessive response validation and introduces a five-level AI Response Validation Framework (ARVF), measured using a six-item Perceived AI Sycophancy Scale (PASS). Across three phases, the study validated the framework with human participants through an online questionnaire, tested LLM-as-evaluators in answering PASS, and examined how ten contemporary LLMs generated and evaluated responses to real-world work and personal conflict scenarios. Results supported the intended progression of perceived sycophancy in ARVF and the reliability of PASS. Trust and perceived competence followed an inverted U-shaped pattern. LLM ratings reproduced the five-level structure but showed calibration differences from human ratings. Based on the LLM ratings, the ten contemporary LLMs were grouped according to their behavioral pattern. Text-based analysis provided further support on the linguistical structure for sycophantic responses generated by LLMs. Findings highlight AI sycophancy as a graded, context-sensitive interactional phenomenon.
Read the full abstract

Portfolio Management

Causal Price-of-Risk Mandates under Overlapping Information

We study whether causal risk mandates constructed from overlapping information blocks can be implemented by one self-financing portfolio that is optimal under pooled information. In an incomplete continuous Brownian market, signal-responsive exposures are projected onto traded Brownian directions and embedded as closed subspaces of a predictable Hilbert space. Each causal mandate is the attainable projection of an identified, baseline-centered intervention surface and is therefore fixed upstream of the allocator. A first exact decomposition separates non-traded structural response, incompatibility across local mandates, and distortion of the common traded book. A fusion operator corrects duplication of common risk directions. The main theorem shows that exact causal decentralization holds if and only if the pooled enlargement preserves reference martingales, the pooled optimum has no component outside the additive desk span, the local causal shadows are compatible with one common book, and that book equals the additive pooled optimum. Under immersion, log-growth regret separates into pooled-interaction loss and common-book distortion, while raw causal implementation error also contains non-attainability and incompatibility. No common benchmark can reduce incompatibility. We derive the unique allocation under a quadratic causal-mandate penalty and exact comparative statics in a two-desk model with shared, desk-specific, and pooled-interaction risk. Additional results cover joint interventions, representation stability, admissible coalitions, and a covariance-weighted continuous-semimartingale formulation. The conclusions are partial-equilibrium statements for continuous markets.
Read the full abstract

2610.00158 · 11 Sep · 28 pages, 3 tables. Complementing earlier work on order-three obstructions to information aggregation (arXiv:2606.26835), this paper develops a framework for causal price-of-risk mandates under overlapping information, focusing on attainability, compatibility, overlap-corrected reconstruction, and implementation by a pooled-optimal self-financing portfolio

Quantum Physics

Symmetry Discovery in Quantum Learning: Observable-Level and Task-Level Inference from Finite Measurements

Symmetry reduces the capacity of a quantum learning model, but the imposed group must match both the measured information and the label transformation. We establish a finite-measurement theory for inferring this group from candidate transformations. The central structural result identifies observable-invisible transformations with the stabilizer of a projected state whenever the probe span is invariant. It turns recovered generators into a valid subgroup and identifies the continuous invisible space with its Lie algebra. For finite dictionaries, an unbiased shadow statistic distinguishes zero from positive squared expectation discrepancies with an inverse-gap measurement rate, improving the inverse-square-gap rate of uniform discrepancy estimation. A commuting qubit lower bound proves the gap dependence optimal at fixed snapshot scale, and simultaneous intervals support data-dependent tolerances. Task validation then tests either the joint distribution through a characteristic kernel or its encoded mean through a classical–quantum discrepancy. An exact group-average identity relates the latter to joint-state asymmetry and specifies its conversion to binary task breaking mass. Projection bias quantifies the cost of excessive symmetry, while an $\ell_1$ readout bound quantifies the capacity gained by relaxing it. At an invariant pure-state backbone, retained and nontrivial breaking sectors are Fisher-orthogonal. Ising-chain calculations connect finite-shot recovery, label-dependent symmetry, and physical sector drift. These results determine which symmetry the measurements support and provide the statistical and geometric basis for a subsequent release decision.
Read the full abstract

Optimization and Control

A Two-Dimensional Extremizer for Relatively Inexact Gradient Descent

We give an explicit extremal function for one step of gradient descent when the gradient error is bounded relative to the true gradient. The function is the sum of a quadratic in one coordinate and a Huber function in the other. It attains, for every admissible relative error level and every stepsize in the intermediate regime, the proposed worst-case constants for both the final squared gradient divided by the initial objective gap and the final squared gradient divided by the objective decrease. This proves Conjecture 2.6 of Vernimmen and Glineur, first stated in June 2025. A change of variable reduces the parameter description to a scalar cubic with a rigorously specified root. We prove all parameter inequalities, derive a nonnegative-square certificate for the upper bound, and verify the matching construction directly. We also characterize the exact one-dimensional worst-case value and prove that two dimensions are necessary in the open intermediate regime.
Read the full abstract

Optimization and Control

Sharpness and Rigidity of the Two-Factor Recht-Re Lower Bound

Products of positive semidefinite matrices arise when successive steps of a randomized optimization method use different data points. Their average can have a negative eigenvalue even though every individual matrix is positive semidefinite. Lai and Lim proved at ICML 2020 that $\sum_{i\ne j}A_iA_j\succeq-\frac{1}{4}n(n-1)I$ whenever $A_i\succeq0$ and $\sum_i A_i\preceq nI$, and conjectured that the constant is sharp for every $n\ge2$. We give an explicit family of $n$ rank-one matrices in dimension $n$ that attains this bound for every $n$, proving that sharpness statement. The construction compresses one direction of an orthonormal basis by a factor $1/\sqrt{2}$. We also give a short, self-contained proof of the known inequality and classify all equality cases: every extremizer contains a common invariant $n$-dimensional block equivalent to the construction. Consequently, dimension $n$ is necessary for equality, and positive definite matrices approach but never attain the lower endpoint. Finally, strongly convex quadratic objectives realize the extremal negative eigenvalue as a two-step expected gradient-descent update.
Read the full abstract

Rings and Algebras

Relative Global Dimension of Controllable Extensions

We study relative global dimensions of extensions $B \subseteq A$ of Artin algebras over a perfect field. A controllable extension is introduced as one for which the relative global dimension is determined by the ordinary global dimension of the quotient $A/AJ(B)A$. General inequalities and sufficient conditions for controllability are established, including the case in which $J(B)$ is a two-sided ideal of $A$. The interaction between relative global dimensions and relative projective dimensions of the regular bimodule is then studied. The class of homologically controllable extensions is introduced, and it is shown to be closed under tensor products. Consequently, for homologically controllable extensions $B \subseteq A$ and $D \subseteq C$, we obtain \[\operatorname{gldim}(A \otimes C, B \otimes D)=\operatorname{gldim}(A,B) + \operatorname{gldim}(C,D).\] The proof relies on a relative Künneth-type result for tensor products of complexes, which may be of independent interest. Finally, controllable extensions with prescribed relative global dimension are constructed, along with families of non-controllable extensions for which the relative global dimension differs arbitrarily from the global dimension of $A/AJ(B)A$.
Read the full abstract

Machine Learning

Product-Aware Deterministic Rounding for Quantized Matrix Multiplication

Scalar rounding decisions interact through matrix multiplication. We study deterministic product-aware rounding after scales, clipping bounds, and grids are fixed, with each active scalar choosing between adjacent levels. For dynamic activation rounding, null-space reduction preserves the relaxed product while leaving at most $r$ fractional decisions, where $r$ is the rank of the active gap-weighted weight block. Conditional-expectation completion gives a deterministic polynomial-time algorithm with squared product error at most $ \mathrm{OPT}_{\mathrm{dyn}}+r\nu_{\max}^2/4$, where $\mathrm{OPT}_{\mathrm{dyn}}$ is the best admissible error and $\nu_{\max}$ is the largest row norm of that block. For reusable static weights, the exact expected product-loss metric is the uncentered input second moment with fixed output bias; free bias recalibration yields the centered covariance. Exact optimization is NP-hard even at rank one. In balanced blocks with $K=1024$ and $r=16$, conditional- expectation completion attains a dither-normalized median error of $0.010$, compared with $0.899$ for round-to-nearest. Clipping-aware initialization reduces median normalized error by a factor of $43.4$ at ten-percent clipping. Held-out Digits experiments show that retaining the input mean or correcting the output bias improves median product error over round-to-nearest in all four tested bit-width and calibration-size settings.
Read the full abstract

All authors Piyush Sao, Narasinga Miniskar, Pedro Valero-Lara, Keita Teranishi, Sudip Seal

Machine Learning

Measure Learning at Steady State: A BIRD-SQL Formula 1 Case Study

Continual Learning Bench scores learning as short-horizon gain versus a reset baseline and finds naive full-context ICL strongest among the memories it tested. We treat ICL as one learning system and score it on a longer shared-world schedule. Steady-state learning is the gap versus baseline on a pre-set late window (last 40 of 174 BIRD-SQL formula-1 questions). We split the score into exploration efficiency (SQL probes), task reward (hits), and delivery cost (API dollars and context size). On gpt-5.6-luna, late probes fall from 4.6-5.6 to 0.95 while hits rise only modestly and ICL context grows to about 95k tokens with cost roughly doubling. Short-horizon gain understates the late probe saving and misses the cost inversion, so we find that unbounded ICL is a poor candidate for the learning mechanism.
Read the full abstract

Machine Learning

When Does Domain Adaptation Help on Physical Vibration Sensors? A Held-Out-Bearing Study of Neural-Operator and Convolutional Models

Diagnosing rolling-element bearing faults from vibration is a canonical physical-sensing task and a widely used benchmark for domain adaptation under operating-condition shift. Accuracies above 99 percent are commonly reported, but under evaluation splits that place the same physical bearing in both training and test. We revisit the task under a held-out-bearing protocol, assigning every bearing unit entirely to either the training or the test set, and find that source-only transfer is far weaker than such numbers suggest: on a change of shaft speed it reaches only $0.36$, against a target-supervised ceiling of 0.97. We then study what governs transfer. Treating computed order tracking, a shaft-angle resampling that places fault frequencies at fixed shaft orders independent of running speed, as a controlled change of representation, we find that a Fourier Neural Operator raises source-only transfer from $0.36$ to $0.61$ on the speed shift, where the fault peaks move, while a convolutional network of matched feature dimension stays near chance in both representations. The representation also decides whether unsupervised alignment can work: with the same normalized RBF-MMD loss and no target labels, the operator reaches 0.71 in the frequency domain but 0.95 in the order domain, within 0.02 of the target-supervised ceiling and above $0.86$ on every held-out bearing fold. Once the representation is right, a small label budget adds little. These results indicate that, for this task, the input representation rather than the alignment method decides whether adaptation helps. A second dataset, whose held-out units are fault diameters rather than bearings, shows that the same protocol exposes failures that even a target-supervised model cannot avoid.
Read the full abstract

Machine Learning

Energy-aware frugal Bayesian optimization

Modern design optimization frameworks aim first and foremost for models with the most accurate predictions without balancing computational overhead. It remains a reason why scaled architecture and multidisciplinary design optimization problems are difficult to address, even with sample-efficient Bayesian optimizers. In this paper, a metric quantifying the computational energy footprint is introduced within a Bayesian optimization framework to guide the parameter setting of a model towards configurations that balance both performance and frugality. The computer experiments highlighted existing tradeoffs between optimum convergence and the underlying energy footprint, and sometimes resulted in both a better-found optimum and lower energy consumption.
Read the full abstract

All authors Gaston Plat, Paul Saves, Nathalie Bartoli, Thierry Lefebvre, Joseph Morlier

Artificial Intelligence

Bringing AI to Autonomous Systems – From Cognition to Collective Intelligence

The purpose of this article is to highlight the central role of autonomous systems as the ultimate stage in the development of AI, to explain the underlying technical challenges that require a combination of connectionist AI and symbolic AI, and to integrate AI and systems engineering. We present a comprehensive framework for the design and evaluation of autonomous systems, based on a generic agent architecture that characterizes their behavior as the composition of cognitive functions organized around a long-term memory containing the agent's evolving knowledge. We address the challenges posed by the implementation of the fundamental features of the agent architecture, in particular the link between sensory data and structured data stored in memory, decision-making related to the achievement of the agent's goals and their planning, as well as the coordination of agents to combine individual and collective intelligence. We explain that agent trustworthiness, unlike that of traditional systems, is not limited to behavioral properties. It includes an essential dimension related to cognitive properties, the validity of which depends on how the agent uses its knowledge in decision-making. We present avenues for the development of methods for evaluating agent trustworthiness. We conclude with a critical assessment of the substantial gap between the aspirational vision of autonomous multi-agent systems and the current state of the art.
Read the full abstract

Game Theory

Sequential Phragmén Guarantees 2-Approximate Core Stability

We study approximately stable committee selection under approval preferences, following Cheng et al.~[EC~2019] and Jiang, Munagala, and Wang~[STOC~2020]. In an approval-based committee election, $n$ voters each approve a subset of the candidates, and a committee of size $k$ is selected. A size-$k$ committee is $\lambda$-stable if, for every nonempty subset $T$ of candidates, the number of voters who strictly prefer $T$ is less than $\lambda$ times the proportional share $\frac{n|T|}{k}$. We consider sequential Phragmén, a classical polynomial-time voting rule, and prove that it always returns a $2$-stable committee. This implies that the $2$-core is always nonempty and improves upon the $3.651$-stability guarantee of Gao, Sun, and Vondrák~[EC~2026]. The guarantee also extends to weighted voters and is tight even for unweighted elections: for every $1\le\lambda<2$, sequential Phragmén can return a committee that is not $\lambda$-stable. Compared with the exact-core algorithm developed in concurrent work by Becker, Greger, and Peters~[arXiv~2026], sequential Phragmén is simpler to implement, and our proof of its $2$-stability is elementary.
Read the full abstract

Commutative Algebra

Algebraic Expressivity Certificates for Shallow Polynomial Neural Networks

We study exact representability by bias-free shallow polynomial neural networks using algebraic geometry. Over $\mathbb{C}$, a width-$r$ network with activation $z\mapsto z^d$ computes a sum of $r$ $d$-th powers of linear forms, whose Zariski closure is a Veronese secant variety. Ideal elimination therefore yields polynomial certificates of nonrepresentability. We implement this construction as a generic architecture-to-certificate pipeline. For quadratics, we recover the exact symmetric determinantal description and explain its dimension through orthogonal symmetry. In higher degree, the implementation recovers classical catalecticant and secant equations and maps the practical reach of direct elimination across a finite architecture sweep. We also derive the exact population loss floor for a rank-two quadratic network on the sphere, illustrating how an algebraic obstruction induces irreducible approximation error.
Read the full abstract

Functional Analysis

Lack of isomorphic embedding from Lorentz function spaces into Lorentz operator ideals

In the present paper, we study the existence of isomorphic embeddings from Lorentz function spaces $L_{p_1,q_1}$ into Lorentz ideals $\mathcal{C}_{p_2,q_2}$ in $B(H).$ In particular, we show that, for $p,q\in (0,\infty),$ the quasi-Banach function space $L_{p,q}(0,1)$ isomorphically embeds into the ideal $\mathcal{C}_{p,q}$ if and only if $(p,q)=(2,2).$ This extends several existing results in the literature and answers a question due to Astashkin et. al.
Read the full abstract

Robotics

D3DWA: Adaptive Weight and Prediction-Horizon for Dynamic Window Approach via Dueling Double Deep Q-Network

The Dynamic Window Approach (DWA) is widely used for local navigation, but its performance depends strongly on parameters that are typically fixed before navigation. In particular, the appropriate prediction horizon can vary with local free space: longer horizons support efficient motion in open areas, whereas shorter horizons help preserve feasible motions in narrow or cluttered regions. This paper proposes D3DWA, an adaptive DWA framework based on a Dueling Double Deep Q-Network (D3QN), which jointly selects the DWA evaluation weights and prediction horizon from a continuous navigation state at every control step while retaining DWA's trajectory generation and collision checking. In eight simulated environments, including unseen layouts, D3DWA reached every goal. Real-robot experiments further showed that D3DWA completed all three tested configurations, including a constrained case in which the weights-only variant timed out. These results demonstrate the benefit of jointly adapting the evaluation weights and prediction horizon. Additional material is available at https://mertcookimg.github.io/d3dwa/
Read the full abstract