Optics
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Thursday 10 September
We present the design of cryogenic metallo-dielectric antennas tailored to single organic emitters, where the choice of host material imposes specific constraints on the antenna geometry. Using dibenzoterrylene (DBT) in para-dichlorobenzene (p-DCB}) as a model system, we show that photon collection efficiencies exceeding 90% can be achieved for arbitrary dipole orientations of the fluorescent molecule. The antenna design is intrinsically broadband and tolerant to emitter positioning within the structure. We further provide concrete fabrication guidelines for the full antenna architecture, and experimentally demonstrate its operating principle by recording a back-focal-plane image (BFP) of a single molecule inside a fabricated antenna.
Information Retrieval
Generative Late-Interaction Embeddings For Visual Document Retrieval
Late-interaction retrieval is the state-of-the-art for visual document search, but it pays for its accuracy in storage. Existing compression methods retain a subset or local average of the N~1,000 vectors per page. Under aggressive storage budgets, however, these methods degrade sharply, and alternatives require retraining the encoder. Investigating this degradation across three encoders, we found two consistent properties: the vectors lie exactly on the unit sphere and concentrate near a manifold of intrinsic dimension five to six. This geometry yields two insights. First, standard k-means centroids fall inside the sphere, causing systematic underestimation of MaxSim scores. Normalizing them to the surface is a free correction worth up to +0.093 nDCG@5 over raw centroids. Second, because the page manifold has few degrees of freedom, the full set of vectors can be regenerated from only a few. To this end, we introduce Generative Late-Interaction Embeddings (GLIE): k << N vectors per page learned from the normalized centroids to serve as both a lightweight index and a basis for regenerating the page's full embedding set. At query time, search runs exclusively on these k vectors, and a decoder expands only the top candidates back to all N vectors for exact rescoring. At four vectors per page on ViDoRe v1, GLIE retains nearly 80% of the uncompressed system's nDCG@5, against 70% for the best prior post-hoc method. These results use a 415K-parameter network fitted in under three GPU-minutes on just a thousand training pages. At a matched training budget, fine-tuning the encoder does not reach even the training-free stage of GLIE, and the full system beats it at every budget. These patterns hold across a second encoder and ViDoRe v2. By reconstructing evidence on demand rather than sampling it, GLIE opens a new axis for storage-efficient retrieval, with the decoder as its main design surface.
Read the full abstract
Machine Learning (Statistics)
Near-Optimal Reinforcement Learning with Multi-Step Transition Lookahead
We study reinforcement learning (RL) with transition look-ahead, where the agent may observe which states would be visited upon playing any sequence of actions before deciding its course of action. Although look-ahead can substantially improve achievable performance, [1] showed that optimal planning with multi-step transition look-ahead is NP-hard. However, this hardness was established using a discount factor close to one. It was therefore unknown whether the problem remains hard for every discount factor, and whether near-optimal planning can nevertheless be performed efficiently. We resolve both questions. First, we show that for every fixed discount factor, exact planning remains NP-hard. Second, we introduce a randomized polynomial-time approximation scheme for every fixed look-ahead depth. Third, we extend our approach to account for unknown transitions. We empirically validate the soundness of our results on the wind-farm storage-control benchmark of [2], showing that our approach, optimally accounting for -step look-ahead information, offers substantially better performance than existing algorithms. [1] Corentin Pla, Hugo Richard, Marc Abeille, Nadav Merlis, Vianney Perchet : On the Hardness of Reinforcement Learning with Transition Look-Ahead [2] Chenbei Lu, Zaiwei Chen, Tongxin Li, Chenye Wu, Adam Wierman : Reinforcement Learning with Imperfect Transition Predictions: A Bellman-Jensen Approach
Read the full abstract
Number Theory
Internal congruences modulo powers of 2 for $k$-colored overpartitions into odd parts
Let $\overline{po}_k(n)$ denote the number of $k$-colored overpartitions of $n$ into odd parts. For every odd number of colors, we establish infinite families of internal congruences modulo high powers of $2$. The proof proceeds by expressing the relevant difference generating functions as a fixed theta quotient multiplied by polynomials in an auxiliary function. Uniform lower bounds for the $2$-adic valuations of the polynomial coefficients then yield the stated internal congruences. These results generalize an internal congruence of Saikia, Sarma and Sellers (J. Ramanujan Math. Soc.~2025) on overpartition triples into odd parts.
Read the full abstract
Human-Computer Interaction
Understanding Operator Attitudes Toward AI-Supported Decision Making in Maritime Operations
Maritime Autonomous Surface Ships (MASS) and AI- supported decision assistants are expected to transform maritime operations, but their safe integration depends on how maritime professionals perceive and trust such systems. This paper presents a survey study on maritime stakeholders' attitudes toward an AI-supported assistant in collision-avoidance scenarios. Participants evaluated technology anxiety, trust in automation, and explanation quality using established and adapted questionnaires, complemented by sentiment and thematic analysis of open-ended responses Results indicate a generally positive disposition toward maritime technology, no clear age-related differences in openness, stable trust across scenarios, and more scenario-sensitive, multidimensional explanation ratings. Open responses showed that participants valued support for decision-making, situation awareness, and confidence-building, while raising concerns about AI reliability, over- reliance and loss of expertise. The findings suggest that maritime AI systems should not focus solely on increasing automation or trust, but on supporting calibrated reliance through transparent, reliable, and operationally meaningful design with domain experts in the loop.
Read the full abstract
Computer Vision
Logit Refiner: Improving Visual Autoregressive Models via Intra-Scale Dependency Modeling
Visual Autoregressive Models (VAR) generate images through next-scale prediction, producing all tokens within each scale in parallel. We show that this parallel decoding constitutes a mean-field-style approximation that discards spatial dependencies among same-scale tokens, causing locally incoherent samples regardless of backbone capacity – a limitation of the decoding rule. Addressing this limitation, we introduce the Logit Refiner, a lightweight autoregressive module that restores intra-scale dependencies by sequentially sampling tokens conditioned on frozen backbone features. Adding only ~10% parameters and less than 5% of the base model's training compute, it plugs into any pretrained VAR checkpoint without retraining. Controlled ablations isolate joint intra-scale sampling – rather than additional capacity or training – as the critical ingredient. Across backbones from 310M to 2B parameters on class-conditional ImageNet 256x256, the refiner consistently improves generation quality, enabling a 1.1B-parameter model to surpass one twice its size. The approach further generalizes to text-to-image generation, confirming that the mean-field bottleneck persists across VAR variants and is effectively alleviated by our method. Project page: https://compvis.github.io/logit-refiner/
Read the full abstract
Geometric Topology
Symplectic Surface Summing and Negative Spheres
We extend the sphere-summing construction of \cite{AZ} from torus sums of spheres to symplectic sums along surfaces of arbitrary genus. The relative surfaces may have arbitrary positive intersection number with the summing surface. We give formulas for the genus and self-intersection of the resulting surface, together with graph and simultaneous versions of the construction. The original elliptic-surface case is recovered as a special case, and higher-genus examples are obtained from hyperelliptic Lefschetz fibrations and braided symplectic surfaces in $\Sigma_h\times S^2$.
Read the full abstract
Combinatorics
The uniform Turán density of the tetrahedron
We prove that the uniform Turán density of $K_4^{(3)}$ is equal to $1/2$. Kielak, Král', Lamaison, Liu, Shu, and Wu have recently proved the same result using combinatorial methods, while our proof is Fourier-analytic. This result answers a question of Erdős and Sós from the founding 1982 paper on this topic.
Read the full abstract
Machine Learning
Thinking with Looped Flows
Humans and machines often solve harder problems by spending more time on computation. In deep learning, looped models implement this idea during inference by recurrently updating a hidden state. In practice, however, their training backpropagates through only one or a few updates, making it hard to train early updates to support future ones. We propose looped flows, an approach that sidesteps this issue by training the recurrence with local denoising objectives. By imposing temporal association across denoising objectives through progressively decreasing noise levels and shared noise, the model is incentivized to learn recurrent states that transfer useful computation over time, even when gradients cover only a few updates. We then formulate inference as integrating the velocity of a probability flow parameterized by the learned denoiser, coupled with recurrent states. This allows solving harder problems by spending more computation through a finer temporal grid and enables multiple valid predictions from different initial noise samples. Across six reasoning benchmarks including two multi-solution benchmarks, looped flows outperform prior state-of-the-art looped models overall, achieving 58.8% test accuracy on ARC-AGI-1 and 12.2% on ARC-AGI-2.
Read the full abstract
High Energy Physics (Theory)
Optimizing the dilaton potential in holographic QCD via phase-transition constraints
In bottom-up holographic QCD (HQCD), two primary approaches are commonly employed: the potential reconstruction method, where the background geometry is fixed a priori and the dilaton potential is derived, and the direct method, where the potential is specified explicitly. While the reconstruction method is highly effective for capturing the phenomenological properties of QCD under extreme conditions, a subtlety arises in that the reconstructed potential depends on the chosen holographic boundary conditions. To address this issue, we propose constructing a dilaton potential designed to reproduce the typical features of HQCD phase transitions obtained by the reconstruction method, rather than one derived solely from HQCD vacuum solution. Using a light-quark model as an example, we construct a potential that captures the principal features of the phase structure of this model. Its parameters are determined at zero chemical potential by minimizing the deviation between the temperature dependence on the horizon position in the direct and reconstructed models and by fitting to the same position of a critical endpoint (CEP). The resulting minimal-potential model satisfactorily recovers the physics of the light-quark model even at relatively small chemical potential. Specifically, it yields a first-order phase transition (FOPT) line close to those obtained from the reconstruction method. Additionally, the minimal-potential model reproduces a rising Cornell-like quark-antiquark potential up to the hadronic distance $\ell=1.5$ fm, which we use as a phenomenological finite-distance criterion for confinement. The location of the confinement/deconfinement crossover in this minimal-potential model is also close to its counterpart in the original light-quark model.
Read the full abstract
Cryptography and Security
SpecGuard: Inference-Time Backdoor Detection For Free
Large language models are often fine-tuned, shared, or downloaded from third parties, so a deployed model may carry a hidden backdoor that behaves normally on benign inputs but switches to attacker-controlled behavior when a secret trigger appears. While backdoors can be audited before deployment, runtime monitoring remains important for models that are frequently updated. The challenge is that LLM serving is latency-sensitive: existing inference-time detectors either rely on assumptions about the trigger form, which can fail on stealthy attacks, or require extra model computation, such as input perturbations or an additional generation pass. We introduce SpecGuard, an inference-time backdoor detector that repurposes speculative decoding at zero added model-computation cost. Speculative decoding speeds up inference by using a small draft model to propose tokens and a target model to verify them. We observe that this verification process already exposes a useful signal: when a backdoor is triggered, the target model shifts toward the attacker's behavior, while a clean draft model does not predict this shift, causing the draft-token acceptance rate to change. We formalize when this signal appears and show that an attacker who suppresses it must also weaken the backdoor. Across diverse backdoor types and model families, SpecGuard reliably detects triggered behavior, including stealthy cases where input-level filters are blind, while avoiding the extra generation cost of existing runtime detectors. Speculative decoding therefore doubles as a free, always-on signal for detecting backdoored LLM behavior.
Read the full abstract
Earth and Planetary Astrophysics
Planetary Accretion Is Less Frequent in Wide Binaries: Evidence from Metal-Enriched White Dwarfs in DESI DR1
Binary stars are common in the Galaxy, and understanding how stellar binarity influences the formation and evolution of planetary systems is an active area of research. In this study, we use metal-enriched white dwarfs in wide binaries as tracers of long-lived planetary systems. With Data Release 1 from the Dark Energy Spectroscopic Instrument (DESI), we find that the fraction of cool metal-enriched white dwarfs in wide binaries is 9.8\,$\pm$\,2.1%, significantly lower (4.7\,$\sigma$) than the 20.5\,$\pm$\,0.9% in a control sample of single systems. Furthermore, we identify a tentative dependence of metal enrichment on projected separation and white dwarf effective temperature, where enrichment fraction decreases at smaller separations and lower temperatures. These findings indicate that, compared to single stars, binary systems either start with smaller initial planetary reservoirs due to suppressed planetesimal formation or undergo more rapid depletion of planetary material during the initial part of the white dwarf stage.
Read the full abstract
Analysis of PDEs
Isolation of scalar Allen-Cahn local minimizers
We show that scalar local minimizers of the Allen-Cahn functional with the double-well potential and homogeneous Neumann boundary conditions on a bounded Lipschitz domain are isolated in the $L^1$-topology. This affirmatively answers a question raised by Kohn and Sternberg (Local minimisers and singular perturbations, Proc. Roy. Soc. Edinburgh Sect. A 111 (1989)). The proof implements an analytic Lyapunov-Schmidt reduction inspired by the theory of monotone dynamical systems in works of Jiang-Yu and Hirsch-Smith in an elliptic setting.
Read the full abstract
Quantum Physics
Absolutely Maximally Entangled States for Odd Prime Power Local Dimension
An absolutely maximally entangled (AME) state is a pure state of $N$ parties with $d$ levels each whose reduced states on every subset of at most half the parties are maximally mixed. Such states provide perfect tensors and optimal quantum error-correcting codes. Determining the values of $N$ and $d$ for which they exist is one of the central problems in multipartite entanglement. For every odd prime power, $q=p^e\ge3$, we construct a stabilizer $\mathrm{AME}(2q,q)$ state and prove the existence of stabilizer $\mathrm{AME}(N,q)$ states for $N=2,3,\ldots,2q$. The construction uses a closed-form $q\times q$ bordered-circulant matrix $A_q$ over the finite field $\mathbb{F}_{q^2}$ to generate a Hermitian self-dual maximum distance separable (MDS) code $[2q,q,q+1]_{q^2}$. These codes lie outside the generalized Reed-Solomon family, including its extended forms. In suitable local bases, the parent states yield Butson Hadamard matrices of order $q^q$ with $p$th-root entries, which become $q$-unitary after normalization. Paired projections on antipodal coordinates give all even lengths from $q+1$ to $2q$; one-party projections and standard short-length constructions complete the range. Weighted submatrices give further explicit realizations.
Read the full abstract
Probability
Small-Ball Marginals Do Not Control Restricted Eigenvalues by Euclidean Gaussian Width
Banerjee, Chen, and Sivakumar asked at COLT 2015 whether a uniform small-ball condition on the rows of a random design matrix forces a restricted-eigenvalue lower bound whose sample complexity is governed by the ordinary Euclidean Gaussian width of an arbitrary spherical subset. We give a negative answer to the natural distribution-free formulation of that question. For every sample size $n$, we construct a centered, genuinely heavy-tailed row distribution in dimension $p=4^n+1$ and a set $A=C\cap\mathbb{S}^{p-1}$, where $C$ is a closed polyhedral convex cone, such that \[ \inf_{v\ne 0}\mathbb{P}\!\left( \left|\left\langle Z,v\right\rangle\right| \ge \frac{\left\lVert v\right\rVert_2}{\sqrt{2}} \right)\ge \frac{1}{12} \quad\text{and}\quad w(A)<2. \] Nevertheless, for the matrix $X$ with $n$ independent copies of $Z$ as rows, \[ \mathbb{P}\!\left( \inf_{u\in A}\left\lVert Xu\right\rVert_2^2=0 \right) \ge 1-\exp(-2^n). \] Thus no positive constants depending only on the fixed small-ball parameters can yield a lower bound of the proposed form $c_1n-c_2w(A)^2$ with high probability. The construction isolates the obstruction: a marginal small-ball lower bound controls every fixed direction, but does not control the distribution-dependent complexity of searching over many directions. We state the quantifiers explicitly and discuss why isotropic or upper-tail assumptions lead to a different, still meaningful problem.
Read the full abstract
Information Theory
Second-Order Expansion of Privacy Amplification Under f-Divergence Criteria
We derive the second-order asymptotics of randomness extraction from memoryless sources with side information under security criteria based on a broad class of Csiszàr f-divergences, treating both a fixed reference side-information marginal and optimization over that marginal. The conditional varentropy decomposes into fluctuations of the conditional entropy across different values of the side information and the average variance of the conditional surprisal for each value. Without marginal optimization, these contributions yield a Gaussian-mixture second-order profile. With marginal optimization, they combine into the total conditional varentropy, yielding a single Gaussian profile. As corollaries, we obtain second-order expansions for Rényi-entropy criteria of all orders $\alpha \in (0,1)$ and recover the known expansion for total variation distance.
Read the full abstract
Mathematical Physics
A closed signature formula for the Katz-Long-Moody Hermitian form
The Katz-Long-Moody construction associates to a representation of the semidirect product of a free group and a braid group, defined by the Artin action, and a nonzero parameter a new representation of the same group. On the pure braid group it corresponds to Haraoka's multiplicative middle convolution for KZ-type equations. For unitary input and a convolution parameter on the unit circle other than one, the construction equips the quotient representation with a canonical non-degenerate invariant Hermitian form. We give a closed formula for its signature in terms of the eigenangles of the input local monodromies, the eigenangles of their ordered product, and the convolution parameter. The formula accounts for the kernel of the form before passage to the quotient and for signature changes at resonant parameters. It determines precisely when the induced form is definite, answering the definiteness problem posed in the companion paper. Definiteness implies unitarizability of the output representation; the converse holds when that representation is irreducible. The proof uses elementary linear algebra: a determinant identity, explicit block-pivot formulas, and an inertia formula for sums of Cayley transforms of unitary matrices. As applications, we compare the rank-one construction explicitly with Haraoka's invariant form for Pochhammer systems, recover the Gauss case of the Beukers-Heckman interlacing criterion, and determine the definite parameter intervals for the Hecke and Temperley-Lieb specializations.
Read the full abstract
Strongly Correlated Electrons
Learning Quantum Matter through Attention in Complex Space
Magnetic many-electron wavefunctions require amplitude and phase to be optimized together. Whether a complex internal representation improves this variational search is a practical question for neural wavefunction design. We introduce Complex Psiformer for interacting electrons in a magnetic moiré continuum, combining complex hidden features and Hermitian-magnitude attention with magnetic boundary conditions and fermionic antisymmetry. After the same number of optimization steps, Complex Psiformer reaches lower energies than Real Psiformer in two finite supercells. Both Psiformers also improve on their respective neural Hartree-Fock references. Across five training seeds in the 25-cell system, the mean Complex advantage is 1.458 meV per electron, with a smaller observed spread. A separately trained two-electron Complex state has a smaller energy gap to a finite configuration interaction reference than its Real counterpart. In the Complex states, flux scans show nonmonotonic density correlations and weaker honeycomb mean-density modulation at higher flux, while connected fluctuations persist. Gauge invariant current maps provide a qualitative comparison of local circulation in the optimized states. These benchmarks support the combined architecture as a variational ansatz for studying energies and charge arrangements in finite magnetic systems.
Read the full abstract
Atomic Physics
Nondipole Coulomb effects in strong x-ray field ionization
Atomic ionization in superintense x-ray fields is studied through numerical solutions of the time-dependent Schrödinger equation including nondipole corrections. The role of the Coulomb potential of the atomic core in the nondipole regime is analyzed for a linearly polarized laser field. In the considered high-frequency and high-intensity stabilization regime, the photoelectron spectra are dominated by a near-zero-energy structure, while above-threshold ionization peaks are suppressed. Two sub-regimes in the nondipole case are identified based on the scaling of the Coulomb momentum transfer, when the Coulomb effect during the interaction with the laser field is either significant or minor. The Coulomb influenced regime sets in within the realm of relatively low field strengths, high laser frequencies, and long pulse durations. An intuitive model explaining the distinct features in the photoelectron momentum distributions and the Coulomb effect scaling is provided.
Read the full abstract
Machine Learning
Dynamic language model representations for multi-objective reaction optimisation
Optimising chemical reactions across multiple objectives, such as yield, selectivity, and safety, is central to chemical synthesis, and model-driven approaches depend critically on how reaction components are represented. Established featurisations are either chemically uninformative, as with one-hot encodings, or, as with molecular descriptors, do not readily extend across chemically distinct components. For structurally and functionally diverse components, it is therefore unclear what a shared representation should contain. Constructing such a representation is itself a challenging research undertaking that must be revisited for each new reaction system. Here we bypass this step by learning the reaction representation dynamically from text. Textual descriptions of reaction conditions are encoded by a fine-tuned language model trained jointly with Gaussian process surrogates, yielding task-adaptive representations within a multi-objective Bayesian optimisation loop. Across nickel- and palladium-catalysed cross-couplings in both sequential and parallel experimentation regimes, this approach reaches optimisation convergence in fewer experiments than descriptor libraries or one-hot encoding. Applied prospectively to a palladium-catalysed cyanation spanning mixed ligand denticity and heterogeneous additives, and to a three-objective asymmetric hydrogenation across chiral iridium and ruthenium catalyst families, two rounds of high-throughput experimentation (192 reactions, under 3% of each design space) delivered conditions translating directly to gram scale in 94% and 84% isolated yield, the latter at 99.6% enantiomeric excess.
Read the full abstract
Formal Languages and Automata
RAGTIMER 1.0: Rapid Rare-Event Partial State Space Construction for Stochastic VAS (extended version)
Transient reachability analysis of rare events in Continuous-Time Stochastic Vector Addition Systems (CTSVAS) such as Chemical Reaction Networks (CRNs) has proven a formidable challenge to cutting-edge tools. Underlying a CTSVAS is a continuous-time Markov chain (CTMC), and CTMC transient reachability analysis calls for Probabilistic Model Checking (PMC). This analysis requires the explicit representation of a model's entire state space. Rare events occur with extremely low probability, compounding the challenge of probabilistic analysis. In CRNs, it is imperative to verify the probability of rare events; even a low concentration of a species can have pathological consequences. This paper presents the RAGTIMER 1.0 tool, which efficiently builds a partial state space for a CTSVAS by enumerating traces to a rare event of interest and expanding them to exploit concurrency and cycles, providing a guaranteed lower bound on the probability of a rare event. Guaranteed lower bounds are particularly useful in synthetic biological applications because they indicate how and when a rare event can be experimentally observed. RAGTIMER is an attractive alternative to existing rare event analysis methods for CTSVAS models. It outperforms existing PMC tools and refutes multiple probability estimates from rare-event stochastic simulation on multiple challenging CRN models. RAGTIMER uses optimized data structures, a simple input format, and memory-safe Rust code to improve the scalability and accessibility of PMC for industry professionals.
Read the full abstract
Optimization and Control
Optimal Recursive Composition and Dyadic Phase Laws for Gradient Descent with Predetermined Stepsizes
Predetermined stepsize schedules featuring carefully chosen long steps have recently been shown to accelerate gradient descent (GD) on smooth convex functions. A prominent class of such schedules is built through recursive composition. In this paper, we characterize the convergence of these optimized recursive schedules, revealing a nonconstant log-periodic modulation across prescribed horizons. Specifically, we prove that for every $N \geq 1$, the optimized ConPP/OBS-S schedules satisfy $$f(x_{N-1})-f_\ast \le \frac{1}{2N^p\Phi(\log_2N)-1} \frac{L}{2}\|x_0-x_\ast\|^2, \qquad p=\log_2(1+\sqrt2), $$ where $\Phi$ is a positive, Lipschitz, nonconstant $1$-periodic function. We derive this by proving that balanced splitting is optimal at every horizon for these constructions, resolving a conjecture of Zhang and Jiang. Furthermore, for their asymmetric counterparts (ConPD/OBS-F schedules), we show that although optimal splits are not necessarily balanced, the same Silver exponent asymptotically persists alongside a distinct log-periodic modulation.
Read the full abstract
High Energy Astrophysical Phenomena
Spectral evolution of NS binary system GX 349+2 using AstroSat observations
We present a broadband spectral study of the Sco-like Z source GX 349+2 using AstroSat LAXPC and SXT observations obtained during 2019 and 2024. The X-ray light curves exhibit large variability, and the hardness-intensity diagram traces the characteristic Z-shaped track. Using hardness ratio and intensity, the Z-track is segmented into six regions comprising normal branch, soft apex, and an extended flaring branch. We perform simultaneous broadband spectral fitting in the energy range 0.7-25.0 keV using two model configurations, tbabs*(thcomp*bbodyrad+diskbb) and tbabs*(bbodyrad+thcomp*diskbb), to study the evolution of emission from the neutron-star boundary layer, accretion disk, and Comptonizing corona along the Z-track. The comparative analysis reveals that, irrespective of the adopted Comptonization geometry, the inferred inner disk radius remains large (~100 km) and exhibits no systematic inward motion as the source luminosity increases. This suggests that the accretion disk remains truncated and the disk luminosity tends to saturate, consistent with a radiation pressure influenced inner disk. The relatively small variation in the inferred mass accretion rate across the branches indicates that changes in $\dot{M}$ alone cannot account for the observed Z-track evolution. Instead, the boundary-layer component becomes progressively dominant toward the flaring branch, with both its temperature and flux increasing with total luminosity, while the relative contribution from the disk decreases. These results indicate a redistribution of accretion power toward the boundary layer and the associated Comptonizing region, suggesting that the primary energy dissipation in GX 349+2 occurs within the boundary layer region of the neutron star.
Read the full abstract
Computation and Language
Beyond Word Error Rate: A Switch Aware Evaluation of ASR and Audio Language Models on English Yoruba Code-Switched Speech
Automatic speech recognition (ASR) systems and audio language models (audio LMs) now report low error rates on monolingual benchmarks, but their behavior on code switched speech in low resource, diacritic rich languages remains poorly characterized. We present a switch aware evaluation of eleven modern systems (six ASR models and five audio LMs) on English Yoruba code-switched speech, using a deterministic 2000 utterance evaluation set and a shared scoring pipeline. Beyond word error rate (WER), we report switch localized diagnostics: a switch entry token error rate (SETER), windowed switch point error rates, language specific error rates, and a diacritic insensitive WER. Our central finding is that aggregate WER hides code switching behavior. The best system by WER (an ASR model) is statistically indistinguishable from a leading audio LM on WER, yet the audio LM is significantly better on every switch localized metric. Across faithful systems, Yoruba token recognition collapses (error 0.97 for almost all systems) while English tokens are recognized far better, and errors concentrate sharply at switches into Yoruba. Several generative audio LMs fail as exact transcribers, producing translation, verbosity, and prompt leakage that are strongly prompt dependent. We release manifests, metric implementations, and evaluation scripts to support reproducible, switch aware benchmarking for African code switched speech.
Read the full abstract
High Energy Physics (Theory)
Logarithmic supertranslations as asymptotic symmetries of gravity at null infinity
Logarithmic supertranslations have been shown recently to be symmetries of the gravitational field at spatial infinity. We extend this work by proving explicitly that they are also symmetries at null infinity. We also show that the logarithmic supertranslation charges match at the “corner" where spatial infinity and null infinity meet. This fully establishes that the asymptotic symmetry group of gravity in the asymptotically flat context contains the logarithmic supertranslations.
Read the full abstract
Optimization and Control
Constant Steps Are s-Composable: An Exact Interpolation Certificate for Gradient Descent
Grimmer, Shu, and Wang asked whether a balanced constant schedule is $s$-composable at every horizon. More precisely, for an integer $n\geq 1$, let $\bar{h}=1+r$, where $r\in(0,1)$ is the unique solution of $$r^n\bigl(1+n(1+r)\bigr)=1,$$ and run gradient descent for $n$ steps with normalized stepsize $\bar{h}$. The cases $n=1,2$ were known, while the general case $n\geq3$ was left open. We prove the conjecture for every $n$. Our proof gives an explicit, dimension-free nonnegative linear combination of the smooth convex interpolation inequalities. The certificate is assembled from matrices supported on contiguous index intervals. Its off-diagonal entries are automatically positive, its quadratic part is diagonal, and its remaining multipliers reduce to two scalar families. We derive closed forms for those families and prove positivity using strict concavity and elementary rational inequalities. Consequently, for every $L$-smooth convex function, the constant schedule satisfies the sharp mixed terminal Lyapunov inequality conjectured in the original paper, together with the associated simultaneous objective-gap and gradient-norm bounds. All exceptional horizons and boundary indices are treated explicitly.
Read the full abstract
Information Theory
Linear Codes over $\mathbb{F}_{q}+u\mathbb{F}_{q}$ associated with Simplicial Complexes, Their Gray Images, and Subfield Codes
In recent years, simplicial complexes have gained considerable attention as a useful tool for constructing distance-optimal codes over finite fields. In this article, we construct four infinite families of linear codes over the ring $\mathcal{R}=\mathbb{F}_{q}+u\mathbb{F}_{q}$ with $u^2=0$ using simplicial complexes with one or two maximal elements, and completely determine their Lee weight distributions via exponential-sum techniques. By employing a Gray map on $\mathcal{R}$, we obtain infinite families of distance-optimal codes over $\mathbb{F}_{q}$, including a near-Griesmer family, and establish sufficient conditions for their minimality. Furthermore, we investigate the corresponding subfield codes and derive sufficient conditions for their distance-optimality and minimality, yielding infinite families of Griesmer and near-Griesmer codes.
Read the full abstract
Functional Analysis
Completeness of constructible norms
In the present work we study for a given Banach space the quotient of the family of equivalent norms modulus the collinearity relationship and prove that under a Banach-Mazur type metric this family is a complete metric space, and we also prove the completeness of the family of constructible norms for the Banach space of bounded Lipschitzian mappings between convex sets of a Banach space.
Read the full abstract
Systems and Control
Predefined-Time Leaderless Consensus Under Denial-of-Service Attacks
This paper addresses predefined-time resilient consensus of leaderless second-order nonlinear multi-agent systems under denial-of-service (DoS) attacks, motivated by coordination requirements in safety-critical applications. The agents are subject to bounded external disturbances and communicate over a strongly connected directed graph whose links are simultaneously disabled during attacks. We develop a switching sliding-mode protocol with the objective of reaching an invariant manifold of position and velocity agreement. The protocol uses relative position and velocity information during attack-free intervals and local velocity feedback during communication blackouts. A time-scaling function remains constant during each blackout and resumes evolving when communication is restored, accounting for the time available for consensus. Under bounds on attack duration and frequency, we derive sufficient gain conditions through a Lyapunov analysis. We show that, despite bounded disturbances, the agents achieve position and velocity consensus by a realistic settling time equal to a prescribed convergence duration plus the cumulative attack duration up to the realistic settling time. The prescribed convergence duration is independent of the initial conditions, and the realistic settling time reduces to that duration in the absence of attacks.
Read the full abstract
Machine Learning
Predicting Privacy Leakage from Weight Spectral Density
Membership inference attacks (MIAs) are widely used to audit the privacy disclosure risk of machine learning models, however current state-of-the-art attacks require training computationally expensive shadow models, making large-scale privacy evaluation impractical. In this work, we investigate whether inexpensive spectral metrics derived from the heavy-tailed self-regularisation framework can serve as proxies for MIA vulnerability. We evaluate several WeightWatcher spectral metrics on image and tabular classification tasks and compare their relationship with MIA privacy leakage against conventional measures of generalisation. Across datasets, stable rank exhibits a strong positive correlation with overall MIA success, while Log alpha-Norm shows a consistent negative correlation with MIA vulnerability at the low false-positive regime. These associations are observed to be stronger than those obtained using the generalisation gap. The results indicate that neural network spectra may contain information about privacy leakage that is not fully captured by conventional measures of overfitting, motivating spectral analysis as a promising direction for scalable privacy auditing.