Superconductivity
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 9 November
Understanding interactions between different low-energy excitations is a central theme in condensed-matter physics. Magnons and superconducting condensates represent two distinct macroscopic quantum states, yet their mutual interaction remains largely unexplored. Here, we demonstrate supercurrent-controlled magnon transport in a ferromagnet/superconductor heterostructure. We observe that the magnon transport signal decreases as the supercurrent increases. This suppression originates from a supercurrent-induced reduction of the superconducting gap, which enhances magnon dissipation into the superconductor via Bogoliubov quasiparticles. Our results provide direct experimental evidence of a supercurrent-controlled interaction between magnons and quasiparticles in the superconductor. The approach used in this study relies on supercurrents, enabling magnon modulation without Joule heating. These findings establish a new route to control magnons via superconductivity and pave the way for energy-efficient magnonic devices.
Robotics
Planners Exploit Deterministic Error: Dithered Low-Bit World Models
Sampling-based planners such as MPPI and CEM query a learned world model thousands of times per control step, so low-bit world models are an obvious way to make model-based control cheaper on a robot. With round-to-nearest activation quantization and no calibration, we find that planning can collapse, by up to 100 points of success, while the world model's latent one-step prediction error stays no larger than that of a stochastic rounding scheme that plans well. At the same bit width, what decides the outcome is not the size of the rounding error but whether it persists across planner iterations and is shared across candidates. Round-to-nearest leaves an error that persists, and an iterating planner drifts toward the candidates that this error favors; in TD-MPC2, holding the same dither fixed within a step or an episode brings the drift back. Dither shared by all candidates and redrawn at every iteration removes most of the damage at the same bit width and without training: at least half of it in all 19 affected Meta-World cells (task $\times$ bit width) of TD-MPC2 (median 95%), in 4 of 5 DINO-WM cells, and in all 8 affected tasks of a 317M-parameter multi-task TD-MPC2. Captured as CUDA graphs on an Ampere GPU, a W8A4 integer TD-MPC2 world model runs its planning rollouts 1.9-3.3$\times$ and the whole planning step 1.3-1.5$\times$ faster than FP32, though not faster than BF16 with pre-converted weights; shared dither adds at most 2.3% latency. In TD-MPC2, block-scaled NVFP4 removes most of the damage and MXFP4 part of it, which ties the failure to coarse grids with a shared scale. Hypotheses, thresholds and analysis code were fixed before the data that test them, and the hypotheses that failed are reported.
Read the full abstract
Nuclear Theory
Collective Signatures of Pionic Correlations at High Momentum Transfer
High-momentum-transfer reactions have provided experimental evidence for enhanced pionic correlations in nuclei, yet the microscopic connection between the low-energy pionic response and the delta-hole excitation spectrum remains unresolved. Here we identify hybrid pion-delta collective modes that emerge through the coherent coupling of nucleon-hole and delta-hole excitations. By evaluating the full delta-hole Lindhard function within the random-phase approximation, we find a robust low-energy hybrid pion-delta mode and a higher-energy collective delta-hole mode. The low-energy mode is strongly enhanced as the Landau-Migdal parameter is reduced, exhibiting behavior characteristic of a precursor to pion condensation. We further show that these modes may be probed through high-momentum-transfer charged-pion production reactions. Because the same collective pion-delta function governs both the response and the RPA correlation energy, our results connect experimentally observable collective excitations with microscopic many-body correlations.
Read the full abstract
Combinatorics
Homomesy of tropical $T$-systems for pairs of simply laced Dynkin diagrams
For a pair of simply laced Dynking diagrams $X$ and $X'$, we prove Chin's homomesy conjecture on the number of red and blue mutations in a period of the tropical Zamolodchikov $T$-system associated to $(X, X')$ using Nakanishi's dilogarithm identity.
Sound
Towards Trustworthy Mobile Speech Interaction: Ensuring Voice Authenticity and Transmission Integrity
The rapid progress of generative speech models has made high-fidelity voice cloning and editing cheap, fast, and widely accessible, weakening the security of voiceprint-based authentication in mobile speech interactions. Existing speech watermarking techniques focus on offline protection or watermarking synthetic speech, and thus cannot verify whether a live utterance originates from the genuine speaker at capture time. We present MoBiMark, a real-time mobile biologic speech watermarking system that binds ultrasonic articulatory signatures to speech content and embeds a dynamic watermark during capture. MoBiMark emits a multi-frequency ultrasonic probe on commodity smartphones, extracts echo features from the vocal tract, and fuses them with a content-dependent speech embedding to produce a dual-layer dynamic watermark combining a persistent user identifier (UID) and a per-utterance speech identifier (SID), enabling both source authentication and propagation-time tamper detection. A jointly trained embedder and detector injects the watermark while preserving perceptual quality, and system-level optimization with INT8 quantization enables lightweight on-device deployment. Evaluated on 50 users across four smartphone models, MoBiMark achieves an average 99.79% accuracy in source and propagation attacks, with negligible impact on speech quality and intelligibility.
Read the full abstract
Analysis of PDEs
Nonexistence of cusp singularities for gravity waves of extreme form
In this paper we revisit the geometric approach developed by Vărvărucă and Weiss for the Stokes conjecture for gravity water waves near stagnation points, in both the irrotational and rotational settings. They introduced monotonicity formulas and frequency formulas to classify all possible singularities of solutions at stagnation points. Two questions were left open: 1. whether cusp singularities can be excluded in higher dimensions $n\geq 3$ [Acta Math. 206 (2011)]; 2. whether cusp singularities for rotational waves can be excluded under the Rayleigh-Taylor condition [Ann. Inst. H. Poincaré C Anal. Non Linéaire 29 (2012)]. We answer both questions in the affirmative. Our main results show that cusp singularities do not exist in higher dimensions and cannot occur for rotational waves even without assuming the Rayleigh-Taylor condition. Our argument combines a delicate analysis of the weighted volume arising from the corresponding free boundary problem with a weighted isoperimetric inequality.
Read the full abstract
General Relativity and Quantum Cosmology
Horizon-penetrating rotating black-hole configurations in VCDM
A regular black-hole metric need not define a regular configuration of VCDM, whose preferred foliation is part of the gravitational solution. We construct stationary rotating, possibly charged configurations on the constant-auxiliary Einstein-Maxwell branch. Their Kerr-Newman and Kerr-Newman-de Sitter geometries carry nonzero constant-mean-curvature foliations that cross the future black-hole horizon and, in the de Sitter case, the future cosmological horizon. A horizon-regular flux formulation specifies the inner data and the distinct hyperboloidal and expanding asymptotic ends. Seventeen survey configurations are supplemented by a mean-flux scan at fixed four-geometry and paired flat/de Sitter comparisons down to $\kappa M = 0.01$, where $K_0 = 3\kappa$. Varying the flux changes the preferred-foliation shear without changing the background seen by standard test matter. The paired local lapse and shear differences have effective powers close to two over the computed range, with the lapse comparison requiring a fixed time normalization. Continuous residuals, geometric reconstructions, and selected grid comparisons test the numerical solutions. An exact lapse bound excludes zeros in the regular stationary exterior block with $\Delta_r > 0$. We distinguish singular finite-proper-distance endpoints from asymptotic trumpet ends and give conditional bounds on a stationary spacelike inner limit. A local cosmological interpretation requires additional matching assumptions; dynamical selection and the rotating interior endpoint are not determined by the stationary construction.
Read the full abstract
Artificial Intelligence
Wireless AI Networks: Co-Adapting Communication Resources and Distributed Intelligence
Distributed artificial intelligence (AI) increasingly relies on wireless networks, where communication and computing resources shape how AI tasks are configured and executed. However, resource availability alone does not determine how effectively these tasks can be supported, since different information selections and transmission strategies may yield different AI performance under the same resource constraints. To address this challenge, this article presents Wireless AI Networks as a unified design perspective for bidirectional adaptation between wireless communication and distributed intelligence. Specifically, an AI-executable resource budget characterizes candidate AI configurations, while task-level information value guides their selection and wireless resource allocation. Building on these principles, a wireless-AI co-adaptation framework coordinates resource abstraction, joint configuration and allocation, and execution feedback. Two complementary experiments illustrate this perspective: update selection and statistical correction improve federated learning under identical model-update payload budgets, while sensitivity-aware power allocation reduces model accuracy degradation under the same average transmit-power constraint. Together, these results highlight the importance of coordinating resource feasibility and task-level information value in distributed AI execution.
Read the full abstract
Differential Geometry
A Cheeger-Müller Theorem For Flat Superconnections
For a flat superconnection ${\mathbb{A}}$ on a ${\mathbb{Z}}_2$-graded vector bundle $E$ over a closed oriented odd-dimensional manifold $M$, we prove a Cheeger-Müller type theorem as proposed by Arias Abad and Schätz. Namely, that the topological Reidemeister norm of ${\mathbb{A}}$ as defined by them equals the Ray-Singer metric defined by Mathai and Wu. The main step proves an analytic reduction result identified by Arias Abad and Schätz and formulated as a conjecture by the author, that the Ray-Singer metric coincides with that of the fibrewise cohomology $H(E,\partial)$ with its Gauss-Manin connection, where $\partial$ is the component of ${\mathbb{A}}$ of form degree zero, under the Knudsen-Mumford isomorphism associated to the form-degree spectral sequence. As an application, we consider Goette's Morse flat superconnection of a fibrewise Morse function on a fibre bundle with even-dimensional fibres over an oriented odd-dimensional base.
Read the full abstract
Strongly Correlated Electrons
Frustrated Magnetism from Dimerized $S = 1/2$ Co$^{2+}$ in Rose-Phase Na$_2$CoP$_2$O$_7$
We report the growth of high-quality single crystals and the magnetic, thermodynamic, and dielectric properties of rose-phase Na$_2$CoP$_2$O$_7$, a low-symmetry magnet in which corner-sharing Co$^{2+}$ dimers form an approximately triangular network. Magnetic measurements reveal dominant antiferromagnetic interactions and pronounced anisotropy among effective $S=1/2$ Co$^{2+}$ moments. Broad low-temperature maxima in the magnetic susceptibility and heat capacity suggest the development of short-range magnetic correlations below approximately 4 K. The large separation between the absolute Curie–Weiss temperatures and the characteristic correlation temperature indicates strong magnetic frustration arising from competing interactions within the coupled-dimer network. Applied magnetic fields suppress these anomalies and reduce the recovered magnetic entropy, suggesting further enhancement of frustration under field. A systematic magnetodielectric response closely follows the field dependence of the magnetization, indicating coupling between magnetic and dielectric degrees of freedom. The low lattice symmetry enables multiple inequivalent exchange pathways and anisotropic interactions while relaxing symmetry constraints on coupling among spin, lattice, and dielectric degrees of freedom. These results establish rose-phase Na$_2$CoP$_2$O$_7$ as a platform for exploring frustrated quantum magnetism and coupled functionalities in effective $S=1/2$ magnets.
Read the full abstract
Artificial Intelligence
ARMOR: Adaptive Curriculum Meta-Learning for Noise-Robust RAG Reasoning
Retrieval-Augmented Generation (RAG) systems have demonstrated remarkable effectiveness in mitigating hallucinations by incorporating external knowledge. However, the retrieval process inevitably introduces noise, posing significant challenges to RAG robustness. Fundamentally, noise robustness is a composite competency encompassing noise identification, retrieval filtering, and robust reasoning. Unfortunately, existing methods often fail to capture this holistic nature, typically restricted to single-dimensional enhancements or hampered by noise-induced optimization interference. To address these limitations, we decompose RAG robustness into three sub-tasks: knowledge boundary perception, retrieval verification, and robust reasoning. Accordingly, we propose ARMOR, an Adaptive Curriculum Meta-Learning for Noise Robust RAG Reasoning. Inspired by the teaching process, ARMOR introduces an adaptive task scheduler as a teacher to curate training plans according to the absolute learning progress of RAG generators (Student). Distinct from GRPO, ARMOR utilizes inter-group advantage estimation during learning to effectively mitigate mode collapse caused by limited output spaces. Experiments on three complex reasoning datasets (2WikiMultiHopQA, MusiQue, and HotPotQA) indicate that ARMOR achieves synergistic improvements in retrieval noise robustness, significantly outperforming baselines in output quality.
Read the full abstract
Computer Vision
Towards Rotation-Only Imaging Geometry: Multi-View Pose Estimation
Global structure-from-motion (SfM) commonly refines relative poses on individual image pairs and then estimates global rotations by rotation averaging. Pairwise refinement does not explicitly combine feature-level constraints with multi-view consistency during rotation initialization. We introduce Global Rotations from Aggregated Matches (GRAM), which estimates global rotations directly from initial relative rotations and inlier matches by minimizing edge-wise rotation-only algebraic residuals. GRAM replaces two-view relative-pose optimization and rotation averaging with a single stage. We further extend rotation- only imaging geometry to multiple views and develop Multi-view Optimization with Rotation-only Reprojection Error (MORE) for nonlinear global rotation refinement. Analytical Rotation-guided Translation (ART) derives locally scaled triplet translations from Linear Global Translation (LiGT) constraints and recovers global translations using weighted least squares. Synthetic and real-world experiments demonstrate the accuracy, robustness, and efficiency of the proposed methods.
Read the full abstract
Optimization and Control
A level-set method for shape and topology optimization of Stokes–Darcy flows with Beavers–Joseph–Saffman coupling
We propose a level-set method for shape and topology optimization of Stokes–Darcy flows with Beavers–Joseph–Saffman interface coupling. We formulate the Stokes–Darcy system with BJS interface conditions using a level-set representation and establish existence and uniqueness of its solution. We further prove existence of minimizers for the viscous-energy minimization and least-squares inverse problems within this framework. Continuous Fréchet differentiability of the control-to-state map is proved, and objective derivatives are derived using an adjoint approach. A reaction–diffusion evolution scheme combined with mixed finite elements is developed to find the optimized interface and accommodate topology changes. Two- and three-dimensional numerical experiments verify the derived sensitivities and demonstrate the effectiveness of the method in minimizing viscous energy dissipation and reconstructing interfaces from partial and noisy velocity observations.
Read the full abstract
Quantum Physics
Super-resolution quantum imaging via multiphoton correlation
The resolution of conventional imaging is fundamentally constrained by the diffraction limit. Here we demonstrate that this limit can be overcome using classical thermal light, through the correlation of multiphoton subsystems. By resolving the photon numbers detected in the two arms of a Hanbury Brown-Twiss interferometer, the correlation function exhibits a narrowed peak or dip whose width is tunable via the average photon number. Temporal and spatial super-resolution enhancements up to 11.8-fold and 5.1-fold, respectively, are experimentally achieved. As a proof-of-principle experiment, a double-pinhole that is completely unresolved by standard ghost imaging becomes clearly distinguishable via the correlation of two-photon versus zero-photon. The principle of this super-resolution imaging can be interpreted as a combination of classical super-resolution imaging and quantum super-resolution imaging. This work provides a versatile platform for super-resolution imaging and lithography, and opens a new avenue for exploring the boundary between classical and nonclassical phenomena.
Read the full abstract
Rings and Algebras
Nonassociative gradings on infinite-dimensional triangular algebras
We study group gradings on triangular algebras relative to a well-ordered basis, with multiplication $x\circ_\lambda y=xy+\lambda yx$. For $\lambda\notin\{0,\pm1\}$, every grading on a finite-dimensional triangular algebra is isomorphic to an elementary grading. In infinite dimension, the same conclusion holds for the algebra of finite-range maps, for the finite-map algebra on a basis of order type $\omega$, and for closed gradings on the full triangular algebra. The infinite-dimensional results also hold for $\lambda=1$ in characteristic different from $2$. For intermediate algebras whose finite-range ideal is graded, we obtain an elementary description relative to a unitriangular basis. The proofs construct neutral associative idempotents with singleton diagonal supports. We also prove that the radical of the full triangular algebra is graded for every nonzero $\lambda$, and that gradings on the finite-range algebra and closed gradings on the full algebra have finite support under the classification hypotheses.
Read the full abstract
Probability
McKean–Vlasov Diffusions from Liouville Spectral Green Response
For Dirichlet Liouville Brownian motion with $0<\gamma<2$, the Gaussian response of a finite vector of logarithmic eigenvalues determines a conditional Green energy. We weight this energy by the current spectral density in the fixed reference space. The resulting self-consistent evolution is a porous-medium equation for the canonical spectral form. We construct its positive probability flow, prove an energy identity, and pass from the regularized response forms to the Green limit. The energy identity yields strong convergence of the Gaussian gradients and of the conditional response drifts. A superposition argument then realizes the flow as a McKean–Vlasov weak diffusion.
Read the full abstract
Materials Science
Strain-dependent Dynamics of Interactions of Iodine-vacancy pairs in CsPbI3: A DFT and Machine-Learning Molecular Dynamics Study
Iodine-vacancy migration and association are important factors in the structural stability of halide perovskites. Although neighboring vacancies have been suggested to form energetically favorable pairs, their stability and reconfiguration under finite-temperature lattice deformation remain unclear. Here, we combine machine-learning molecular dynamics (MLMD) and density functional theory (DFT) to investigate neighboring iodine vacancies in tilted CsPbI3 under isotropic and biaxial strain. The MLMD trajectories reveal a nonmonotonic dependence of short-time hopping fractions on strain, with strong isotropic compression favoring point-like separation and tension promoting pair-preserving relay hopping. Notably, biaxial compression produces distinct orientation-dependent responses: vacancy pairs initially aligned with the compressed plane preferentially undergo separation, whereas those perpendicular to the plane predominantly exhibit relay hopping. Complementary static calculations using DFT and the SevenNet-0 potential yield repulsive interactions for all configurations examined. These findings highlight the importance of finite-temperature dynamics and strain-dependent local geometry in understanding vacancy-pair stability and reconfiguration in halide perovskites.
Read the full abstract
Numerical Analysis
An asymptotic preserving scheme for kinetic equations describing bacterial traveling pulses
We consider a kinetic model for bacterial chemotaxis that describes the propagation of traveling pulses driven by a nutrient and a self-produced chemoattractant. In the diffusive regime, the model converges to a macroscopic drift-diffusion system with bounded chemotactic fluxes. We construct a fully discrete asymptotic preserving (AP) finite difference scheme based on a micro-macro decomposition of the kinetic equation. The method is stable uniformly with respect to the scaling parameter and, as this parameter tends to zero, reduces to a consistent discretization of the limiting macroscopic equations. We prove positivity and mass conservation for the limiting scheme, and use a positivity-preserving correction for the kinetic solver. Numerical simulations reproduce the experimentally observed traveling pulses at both the kinetic and the macroscopic level, validate the theoretical traveling wave speed, and illustrate the convergence of the kinetic model to its macroscopic limit.
Read the full abstract
Computer Vision
LangDepth: Vision-Language Scene Priors for Semantic Binning and Cross-Domain Monocular Depth Estimation
Monocular depth estimation has advanced through large-scale pretraining and strong visual backbones, yet metric depth remains sensitive to scene-level scale. Indoor rooms and outdoor driving scenes occupy very different depth ranges, but most depth decoders infer this global structure implicitly from visual features. We ask whether scene-level semantic structure can instead act as an upstream prior that shapes the metric output space before dense prediction. We present LangDepth (SemanticBins), a monocular depth architecture that couples strong visual features with a scene-conditioned output space. A frozen CLIP encoder and a learnable bank of scene- and geometry-oriented concept keys convert image features into a compact scene descriptor, without requiring per-image text or an external captioner. A Scene Prior Module (SPM) maps this descriptor to a scene-conditioned metric depth interval, providing a global range anchor before dense decoding. In parallel, a DINOv2-L backbone with a lightweight FPN extracts dense visual features, and a Per-Pixel Assignment Network (PPAN) assigns pixels to the scene-prescribed bin structure. A confidence gate fuses the dense visual estimate with the SPM prior. We train LangDepth jointly on NYU Depth V2 and KITTI under a modest supervised depth budget, without web-scale depth supervision or pseudo-labelled depth data. A single model achieves delta_1 = 0.960 on NYU, 0.960 on KITTI, 0.887 on DIODE indoor, 0.946 on iBims-1, and 0.861 on SUN RGB-D. Beyond aggregate accuracy, the SPM predicts compact indoor ranges and much wider outdoor KITTI ranges, the gate becomes SPM-dominant on outdoor scenes, and perturbing the routed scene context causes large drops on KITTI and iBims-1. These results support scene-conditioned output spaces as a practical and interpretable mechanism for cross-domain monocular depth estimation.
Read the full abstract
Statistical Methodology
Optimal Radial Weighting for Robust Mutual Fund Selection
Identifying skilled mutual funds requires methods that distinguish positive alphas from sampling variation when many funds are evaluated simultaneously. Heavy-tailed returns and dependence across funds complicate this task. We propose a robust selection procedure based on radially weighted alpha estimation in linear factor pricing models. The method uses both the directions and lengths of standardized residuals, estimates marginal scales from pairwise residual differences, and accounts for the estimation of factor loadings. Within the proposed class, inverse-norm weighting minimizes the asymptotic variance of the alpha estimators and maximizes the local asymptotic power of individual tests at a common significance level. For radial weights fixed conditional on the factor design, we establish asymptotic false discovery rate control under elliptically symmetric errors and show that the proportion of sufficiently strong positive-alpha funds selected tends to one. The theory accommodates comparable cross-sectional and time-series dimensions and nonvanishing alphas. We also develop a smoothed implementation and assess its finite-sample properties. Simulations show higher true discovery proportions than least-squares and spatial-sign methods under heavy-tailed distributions and improved numerical stability from smoothing. An application to U.S. equity mutual funds examines the subsequent performance of portfolios formed from the selected funds.
Read the full abstract
Computation and Language
BridgeKV: Large-Model Prefill and Small-Model Decoding for LLM Inference via KV Cache Translation
LLM inference has two phases with different bottlenecks: LLM prefill runs once per request and is compute-intensive. Decoding repeats for every output token and is limited by memory bandwidth. To address the challenges associated with this issue, different solutions have been proposed. All of these solutions use the same model for both prefill and decode phases. We propose BridgeKV, which uses a larger model for prefill and a smaller model of the same family for decoding. A lightweight KV-stitching module translates the larger model's prompt cache into the smaller model's cache layout. It handles differences in cache shape and representation through learned local layer mixing and separate nonlinear maps for keys and values. With both models frozen, we train the module to align their caches and make the smaller model follow the larger model's next-token predictions. A second training stage adds low-rank adapters when further task adaptation is needed. Across six model pairs and five datasets, BridgeKV outperforms fine-tuned small-model baselines with matched trainable parameter budgets in all 30 settings. It closes up to 80% of the quality gap to using the larger model for both phases. On one A40 GPU, Llama-3.1-Instruct-8B to Llama-3.2-Instruct-1B achieves 3.0X faster requests than the larger model alone, with latency close to the smaller model alone. BridgeKV can improve the hardware and system solutions that separate prefill and decode phases across GPUs. A separate serving simulation models request arrivals, queues, and cache transfer. For Llama-3.1-Instruct-70B to Llama-3.2-Instruct-1B, it predicts up to 33.86X faster requests and 27.5X lower energy use than using the larger model for both prefill and decode phases.
Read the full abstract
High Energy Physics (Theory)
Proof of the Rényi QNEC
We give a general proof of the Rényi QNEC formulated using the Araki-Masuda weighted $L_p$ norms for all $p > 2$. This generalizes the algebraic proof of the QNEC applicable at $p=2$. Our approach utilizes a weighted version of a conditional $L_p$ norm which extrapolates the quantum expansion/ant-quantity. We elucidate some properties of the corresponding weighted conditional entropy both for half-sided modular inclusions and for general von Neumann subalgebras with a common cyclic and separating vector. The $p=2$ limit computes a conditional dynamical entropy originally defined in the context of black hole dynamics.
Read the full abstract
Quantum Physics
An optimal quantum strategy for the $2$-hat game
We study the two-player hat game as a two-input, three-output non-local game. We calculate the quantum commuting value and show that it coincides with the quantum one. In particular, we show that \[\omega_{\rm q}(H_2)=\omega_{\mathrm{qc}}(H_2)=\frac{15+\sqrt{33}}{36}\approx 0.576238.\] We achieve the optimal quantum advantage using a shared non-maximally entangled qutrit state. To establish optimality, we exploit the game's symmetry under the dihedral group $D_8$ and apply its representation theory to substantially reduce the complexity of the underlying matrix problem. We also determine the no-signalling value $\omega_{\mathrm{ns}}(H_2)=2/3$.
Read the full abstract
Dynamical Systems
Schmidt games for nonautonomous Liverani–Saussol–Vaienti maps
Liverani–Saussol–Vaienti (LSV) maps form a family of piecewise differentiable dynamical systems on $[0,1]$ depending on a parameter $a>0$. These maps are expanding away from a neutral fixed point. In this paper, we study Schmidt games for nonautonomous compositions of LSV maps with arbitrary parameter sequences in a compact subinterval of $(0,\infty)$. For every countable set $P\subset[0,1]$, we prove that the set of points whose orbit closure is disjoint from $P$ is absolute winning on $[0,1]$. Our proof combines a potential game argument with an interval tree constructed from successive returns.
Read the full abstract
Computer Vision
Calibrated Alarm Thresholds for Compact Fire and Smoke Detectors
Compact object detectors such as YOLOv8n are a common choice for camera-based fire and smoke detection on low-power devices. These detectors are usually evaluated with box-level metrics such as mean average precision, and their outputs are turned into decisions with a fixed confidence threshold such as 0.25 or 0.5. A fire monitoring system, however, makes one decision per image: raise an alarm or not. In this paper we evaluate two publicly available compact detectors on the D-Fire test set at the image level. We compare three simple rules for turning box detections into one alarm score per image, and we use a split conformal procedure to choose the alarm threshold from held-out images so that the expected fraction of missed fire and smoke images stays below a chosen target. We find that the same fixed threshold leads to very different miss rates for the two detectors, 17.3 and 8.4 percent at a threshold of 0.5, while the calibrated threshold meets miss-rate targets of 10, 5 and 1 percent for both detectors and all three rules. At a 5 percent target, the false alarm rate on images without fire or smoke stays between 1.2 and 3.0 percent. We also find that smoke-only images stay below the target even when the overall target is met, and we describe the main failure cases.
Read the full abstract
Databases
Causal Resolution of Undefined Answers in Datalog with Negation
Unstratified negation may leave the answer to a Boolean Datalog query undefined under the well-founded semantics. We study database causality in this setting by asking which fact insertions and deletions realize a prescribed transition and which other changes preserve the observed query answer while making a designated fact pivotal. Such a set of changes to other facts is a contingency for the designated fact and the prescribed transition. Our game representation assigns one shared move to each mutable fact and preserves query values and intervention distances. Pairing source and target certificates characterizes all minimum contingencies for the six transitions among true, false, and undefined. When inserting or deleting a move resolves a draw, a directed tree traces the newly winning and losing positions along alternating paths to the source of that move. For the fixed one-rule win–move program, deciding whether a contingency of size at most $k$ exists is $\mathrm{NP}$-complete for each transition, even on bounded-degree graphs. Computing the minimum size is $\mathrm{FP}^{\mathrm{NP}[\log]}$-complete, and counting minimum contingencies is $\#\mathrm{P}$-hard under polynomial-time Turing reductions. When only one of true and false is reachable, the prime implicants of undefinedness satisfied at the observation cover every undefined state, and their supports induce nonempty subtrees of a common tree, we compute a minimum resolving intervention in polynomial time from the game instance. From the explicit tree and these implicants, we construct an equally large family of disjoint supports, certifying optimality.
Read the full abstract
Robotics
Graph-Sparse Distributed Predictive Control for Multi-Robot Motion Planning vianEdge-Consensus ADMM
This paper presents a distributed model predictive control (MPC) architecture for collision-avoiding motion planning of planar robots with double-integrator dynamics. To reduce coordination demands, a time-varying predicted conflict graph selects relevant pairwise interactions. The associated collision-avoidance constraints are convexified once per MPC update using warm-start trajectories, yielding a fixed convex quadratic program (QP). The resulting QP is then decomposed using an edge-consensus formulation of the alternating direction method of multipliers (ADMM), enabling parallel robot-local optimization and independent closed-form projections onto separating half-spaces. Execution is selectively supervised by a centralized high-order control barrier function (HOCBF) QP when the distributed candidate does not satisfy the direct-execution conditions. Evaluated in C++ simulations involving 8, 16, and 32 robots, the proposed architecture achieves collision-free execution and substantially improves computational scalability while maintaining task-level performance comparable to a centralized complete-graph baseline.
Read the full abstract
Quantum Physics
Advantage Distillation in Paired-State Quantum Key Distribution
We investigate advantage distillation for paired-state Loop-Back quantum key distribution, with a fixed optical station serving a lightweight portable terminal. Alice prepares two polarization-encoded pulses, Bob selects and applies a polarization filter to each pulse, and Alice infers his selections from the returned detection pattern. The proposed advantage-distillation step selects consistent blocks of three two-bit records using two independent public parity checks on a minimal connected graph. The closing edge of a parity polygon is shown to be redundant: any spanning tree of the same vertices has the same acceptance set. We derive exact acceptance probabilities, residual-error laws, and retention rates. Under independent bit errors, the three-record filter reduces the leading error term from e to e^2, while its cost in discarded records and public information is accounted for explicitly. A modular security theorem bounds the additional guessing advantage available to an adversary after syndrome disclosure and block acceptance, including arbitrary quantum side information. We also evaluate a specified single-photon passive-tap optical channel, obtaining a positive conditional entropy bound and exact accepted-block security quantities. This parity-based advantage-distillation step detects discrepancies by postselection; it does not correct individual errors and is intended to precede conventional reconciliation. The architecture suggests two further directions: hybrid filtering and reconciliation, and key establishment between portable users accessing a shared quantum station.
Read the full abstract
Computer Vision
SpinSCI: Mask-Conditioned Spatial-Temporal Spin Transformer for Video Snapshot Compressive Imaging
Video snapshot compressive imaging (SCI) recovers multiple frames from one spatiotemporally coded measurement. This paper proposes SpinSCI, a mask-conditioned Spatial-Temporal Spin Transformer whose Mask-Conditioned Ising Refinement Block (MC-IRB) combines attention logits and sensing information to form signed spin couplings. Finite mean-field updates gate residual feature refinement. Physics-guided initialization and final data consistency incorporate the measurement model. Compared with HiSViT$_9$, SpinSCI achieves average PSNR gains of 0.51 dB, 0.23 dB, and 0.04 dB on small-scale gray, mid-scale color, and large-scale color benchmarks, respectively, with 16.9% fewer parameters, 21.0-21.6% fewer FLOPs, 17.1-49.9% lower latency, and 62.3-63.3% lower peak memory. Relative to DU-9stg, average PSNR differs by at most 0.13 dB across these benchmarks, with 28.2-28.7% fewer FLOPs, 64.6-71.7% lower inference latency, and 30.1-35.0% lower peak memory. Qualitative comparisons on the real benchmark further show recovery of fine details and textures. Code and reference checkpoints are available at https://github.com/skethz/SpinSCI.
Read the full abstract
Machine Learning
Spectral Rank and Step-Direction Dynamics in Language-Model Post-Training
Language-model post-training modifies weights through cumulative updates and successive displacements, yet existing studies report seemingly conflicting spectral properties: nearly full-rank updates and low stable-rank changes. We show these observations are compatible and characterize update directions during training. For singular values following $\sigma_i \propto i^{-\alpha}$ with $0 < \alpha \leq 1/2$, we prove that stable rank becomes negligible relative to fixed-energy rank as dimension increases. Full singular value decompositions of selected OLMo-2-7B updates confirm this separation: three proximal policy optimization (PPO) attention updates have stable ranks of 9.6-18.3 but require 532-1,222 singular components to capture 80% of their energy. We next compute consecutive checkpoint-displacement cosines from full-model inner products. Across nine publicly released reinforcement learning from verifiable rewards (RLVR) stages, all 131 adjacent pairs have negative cosines on their native checkpoint grids, including PPO and Group Relative Policy Optimization (GRPO) trajectories. Supervised fine-tuning comparisons across four model families show configuration-dependent alignment, including negative late-training cosines for Mistral and FP32 Llama. Finally, a controlled OLMo-2-1B study with three paired seeds and a 2,048-token response cap compares group-relative advantages against a learned critic with generalized advantage estimation (GAE). At matched 60-rollout windows, late-training mean cosines are +0.259 and +0.012, respectively; the paired critic-minus-group difference is -0.247 (seed-level 95% interval [-0.358, -0.137]). These findings distinguish spectral concentration, energy coverage, and trajectory direction, and demonstrate that the advantage system influences directional alignment under matched training conditions.