Peer Revenue Returns subject to peer review

Papers

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

All of arXiv

Monday 31 August

Solar and Stellar Astrophysics

Astrometric modeling of unresolved variable binary systems. II. Application to Gaia epoch astrometry of nearby pulsating and convective red giants

(Abbreviated) The interpretation of high-precision astrometry for intrinsically variable stars remains challenging, particularly for unresolved binary systems containing asymptotic giant branch (AGB) stars. In such systems, large-amplitude pulsations and evolving convective surface structures induce time-dependent photocentre displacements that perturb the observed orbital motion. At the same time, Gaia DR4 epoch astrometry offers the prospect of deriving accurate orbital solutions and parallaxes for nearby unresolved AGB binaries. Building upon the variability-induced mover (VIM) framework, we extend the astrometric model for unresolved evolved binaries by combining Keplerian photocentre motion, pulsation-induced flux-dependent photocentre shifts, and convection-induced photocentre motion. In the forward simulations, convection is represented either by stochastic photocentre displacements drawn from an exponential correlation function or by photocentre time series extracted from 3D radiation-hydrodynamic simulations. We then test a retrieval framework that fits the orbital and VIM signal while incorporating a red-noise covariance matrix to account for correlated astrometric residuals produced by convection. We show that Gaia epoch astrometry can recover reliable orbital and astrometric parameters for unresolved AGB binaries despite strong pulsation- and convection-induced photocentre variability. Pulsation-induced variability produces a coherent VIM signal that can be modeled jointly with the Keplerian photocentre orbit, while convection-induced photocentre motion behaves primarily as temporally correlated astrometric noise. Retrievals that ignore this correlated component lead to biased proper motions and parallaxes, whereas the inclusion of a physically motivated red-noise covariance model enables accurate recovery of the underlying orbital solution and astrometric parameters.
Read the full abstract

All authors L. Decin, K. Sivkova, P. Kervella, A. Chiavassa, E. Beguin

Machine Learning

Three Steps at a Time: Learning Representations from Action Sequences in Contrastive RL

While self-supervised approaches to reinforcement learning have achieved strong results by learning representations of states and actions, a key open question is the time scale over which actions should be modeled. Departing from the standard formulation relying on single-step actions, we extend contrastive reinforcement learning (CRL), a prototypical self-supervised method, to operate over action chunks, and find that this results in large, pervasive gains across established offline and online benchmarks: +31.7% and +93.1% across 18 and 11 environments respectively. While action-chunking-driven gains are generally explained through the ability to model non-Markovian, temporally extended policies, and to propagate unbiased multi-step returns, interestingly, we find that these arguments only partially apply to CRL. Our empirical studies suggest that, in the context of CRL, an action chunk carries more information about the goal than a single action, measurably improving the critic's representations, and rendering the algorithm significantly more effective.
Read the full abstract

All authors Michal Korniak, Kamil Dybek, Benjamin Eysenbach, Marco Bagatella, Michał Bortkiewicz

Signal Processing

Energy Efficiency in Microwave Linear Analog Computer (MiLAC)-Enabled Communications

Microwave linear analog computers (MiLACs) have emerged as a promising architecture for energy-efficient wireless communications by shifting signal processing from the digital to the analog domain using tunable impedance networks. Several MiLAC architectures have recently been proposed, including the single-layer MiLAC (SLM), two-layer MiLAC (TLM), and hybrid digital-MiLAC (HDM). In this paper, we investigate the energy efficiency (EE) of these architectures. Specifically, we formulate EE maximization problems for the SLM, TLM, and HDM architectures under transmit power, user rate, and architecture-specific constraints, and develop a dimensionality reduction technique with successive convex approximation (SCA)-based algorithms to efficiently solve the resulting non-convex problems. We further derive a computationally efficient solution for EE maximization based on a search over only $(K+1)$ closed-form candidate solutions, where $K$ is the number of users, and analyze the asymptotic EE in the large-antenna regime under negligible quantization noise. Our analysis shows that the EE of the SLM, TLM, and HDM architectures scales as $\ln(N)/N^2$, whereas conventional digital beamforming (DBF) scales as $\ln(N)/N$, where $N$ denotes the number of transmit antennas. Despite these different scaling laws, the faster EE decay of MiLAC-based architectures becomes relevant only at very large antenna dimensions, typically involving thousands of antennas, while they maintain superior EE over practically relevant finite antenna regimes. Finally, we derive an approximation of the number of antennas required to achieve the maximum EE for each architecture and the corresponding EE. Simulation results show that, over practically relevant antenna regimes, MiLAC-based architectures achieve substantially higher EE than conventional fully digital and hybrid analog-digital beamforming.
Read the full abstract

Differential Geometry

On compact steady pluriclosed soliton surfaces

We complete the classification of non-Kähler steady pluriclosed soliton on compact complex surfaces, as initiated by Streets. The underlying complex structure must be a Hopf surface, with any finite primary cover being of class 1. Moreover, the soliton metric is unique up to biholomorphism.
Read the full abstract

Number Theory

Unit Indices of Shanks Orders

For an integer $t\geq-1$, let $\theta_t$ be the largest real root of $g_t(X)=X^3-tX^2-(t+3)X-1$, and set $R_t=\mathbb{Z}[\theta_t]\subseteq\mathcal{O}_t=\mathcal{O}_{\mathbb{Q}(\theta_t)}$, $N_t=[\mathcal{O}_t:R_t]$, and $\varepsilon_t=[\mathcal{O}_t^\times:R_t^\times]$. We determine $\varepsilon_t$ when $N_t$ is squarefree, $27$, or $343$: the only nontrivial indices in these cases are $\varepsilon_3=3$, $\varepsilon_5=7$, $\varepsilon_{12}=13$, and $\varepsilon_{54}=19$. Local conductor calculations give $\varepsilon_t\mid N_t$ for squarefree $N_t$. For arbitrary $N_t$, a regulator comparison shows that $t\geq2N_t$ implies $\varepsilon_t=1$. When $N_t=p^3$ with $p\neq3$ a rational prime, this bound and the index criterion leave at most four possible parameters with $\varepsilon_t>1$ for each fixed $p$. For arbitrary additive index, we also prove that $13\mid\varepsilon_t$ if and only if $t=12$ or $66$, with unit index $13$ in both cases. The proof determines the rational solutions of a plane quartic equation by an explicit genus-two descent and a local Chabauty argument, proving Louboutin's Conjecture 19 on its integral solutions. When $N_t$ is a rational prime, we determine the Picard kernel, the cardinality and fibers of the ideal class monoid over $\operatorname{Pic}(\mathcal{O}_t)$, and the corresponding integral matrix-conjugacy classes.
Read the full abstract

Machine Learning

MolLedger: An Additive Graph Neural Network with Chemically Grounded ADME Attributions

Optimizing absorption, distribution, metabolism, and excretion (ADME) is an important part of small molecule drug discovery. Many machine learning models have been built to predict ADME properties to facilitate this optimization process, but explaining model predictions is challenging. We propose a new graph neural network architecture with built-in atom attributions. Our model MolLedger learns a global context vector for each molecule and a per-atom head to output atom scores that sum to the predicted property. The atom scores are regularized to align with relevant chemical properties. We prove that MolLedger is a universal approximator and demonstrate empirically that the new architecture obtains explainability with little effect on performance. We show that the interpretations from MolLedger are faithful, concordant with held-out physical properties, and align with the changes between matched molecular pairs. Our case studies comparing interpretations from multiple methods on molecular pairs reveal that MolLedger is much better at producing sensible explanations for predicted property changes.
Read the full abstract

Game Theory

Test-time Reinforcement Learning in Imperfect Information Games

Test-time reasoning has significantly improved performance in domains ranging from games to language models. However, test-time policy changes with formal guarantees on the performance of the resulting strategy remain a challenge in two-player zero-sum imperfect-information games. Existing solutions are limited to tabular methods or single gradient step updates. In this work, we investigate policy-gradient algorithms as a method for scalable test-time reasoning. We extend the concept of gadget game, tabular technique for test-time search, to the reinforcement learning setting. Unlike prior approaches, we represent the gadget game implicitly by modified sampling and neural policy rather then explicitly by constructing it, thereby removing constraints on subgame size. Furthermore, we formally prove that, unlike prior tabular algorithms, regularized policy-gradient algorithms limit possible strategy degradation caused by test-time reasoning, even without the gadget games. Our evaluation across small- and large-scale games confirms that additional test-time training often substantially improves performance relative to the blueprint strategy.
Read the full abstract

Complex Variables

On the Riemann Boundary Value Problem for Poly- and Meta-hyperanalytic Function Spaces over d-summable Curves

Hyperanalytic functions, in the sense established by the mathematician Avron Douglis, are Douglis algebra-valued functions defined via a hypercomplex structure rather than the standard Cauchy-Riemann equations characteristic of traditional complex analysis. The classes of polyhyperanalytic and meta-hyperanalytic functions represent advanced generalizations of Douglis's analysis. They are employed in the study of partial differential equations and elasticity, extending the concept of the classical holomorphic function through higher-order iterations and non-homogeneous terms. The aim of this work is to find solvability conditions for a fundamental Riemann-type boundary value problem for spaces of poly-hyperanalytic and meta-hyperanalytic functions defined on an open, bounded, simply connected subset of the complex plane, where the boundary need only be a closed d-summable curve. In fractal geometry, d-summability is a geometric property used to define the boundaries of fractal domains, enabling advanced mathematical integration and calculus on complex structures defined by Jenny Harrison and Alec Norton.
Read the full abstract

Computer Vision

Quantum-Grassmann-Plucker Token Mixing for Deep Learning-Based Post-Disaster Damage Assessment

Timely post-disaster building damage assessment from satellite imagery is a critical engineering decision support task, yet it remains constrained by class imbalance, ambiguous intermediate damage states, and limited cross-event transferability. This study presents, to our knowledge, the first application of Grassmann-Plucker (GP) token mixing to computer vision and introduces two extensions for image classification: the Quantum-inspired Grassmann-Plucker (QGP) head and the Hybrid Quantum Machine Learning Grassmann-Plucker (HQML-GP) head. The GP head represents multiscale relationships among image patch tokens by encoding subspaces formed by token pairs with Plucker coordinates; QGP enriches these coordinates with amplitude-derived probability features, whereas HQML-GP incorporates expectation values generated by a simulated quantum circuit into the geometric token representation. Paired pre- and post-event image patches from the xBD tornado dataset were processed using a frozen six-channel Vision Transformer base encoder with 16 x 16-pixel patches. The three GP-based heads were compared with multilayer perceptron and Transformer baselines under identical training, checkpoint selection, and evaluation protocols. Joplin and Moore tornado samples were used for model development and seen-event testing, while Tuscaloosa was reserved for unseen-event evaluation. QGP led both test sets in accuracy and macro-F1: 83.46% and 64.50% for the seen events, and 66.45% and 52.70% for the unseen event. Although HQML-GP obtained the highest validation macro-F1 of 65.63%, it did not surpass QGP on either test set and required substantially more training time per epoch. These results establish GP token mixing as a competitive attention-free alternative to conventional Transformer-based token mixing for paired satellite image damage classification.
Read the full abstract

All authors Kooroush Farahkhah, Umut Lagap, Taha Rezaei, Saman Ghaffarian

Computation and Language

GMTS: Gradient Magnitude-based Token Selection Improves RLVR Training for LLM Reasoning

Reinforcement learning (RL), particularly RL with Verifiable Rewards (RLVR), has recently emerged as a central paradigm for enhancing large language models' (LLMs) reasoning abilities, demonstrating remarkable effectiveness across reasoning tasks. Recent studies suggest that high-entropy tokens play an exceptionally important role in model training, since training with only the highest 20% entropy tokens yields significant performance gains. However, why such high-entropy tokens are beneficial remains insufficiently understood. In this work, we find that although high-entropy tokens within one answer tend to correlate with large gradient magnitude, entropy alone fails to consistently reflect token importance across different answers, considering the variations in the answer-level reward signals. Based on this observation, we introduce the Gradient Magnitude-based Token Selection (GMTS) method to quantify token importance, which leverages the entropy-gradient connection to approximate gradient-magnitude rankings for token selection. We find that training on the top 20% tokens ranked by GMTS consistently outperforms entropy-based token selection across three reasoning domains and various model sizes, suggesting that GMTS provides a more fine-grained estimate of token contribution for RLVR training.
Read the full abstract

Quantum Physics

Deterministic Universal Logical Gates for Finite-Energy GKP Qubits in Trapped Neutral Atoms

We present a universal logical gate set for finite-energy Gottesman-Kitaev-Preskill (GKP) qubits encoded in the harmonic motional states of trapped neutral atoms. Internal electronic states serve as ancilla for implementing state-dependent conditional displacements in phase space, enabling arbitrary single-qubit phase gates. A Transient Rydberg excitation-mediated atomic dipole-dipole interaction enables a controlled-Z gate without invoking the blockade mechanism. The gates operate under magic-trapping conditions, allowing continuous trapping throughout the gate sequence, thereby preserving the motional encoding without intermediate measurement or feedforward. We assess the experimental feasibility of the proposed protocols using neutral $^{88}\mathrm{Sr}$ atoms and obtain average gate fidelities of $0.986$ for single-qubit phase gates and $0.985$ for the controlled-Z gate at a finite squeezing parameter of $\Delta=0.25$ (12 dB of squeezing), increasing to $0.997$ and $0.995$ respectively as $\Delta$ reaches $0.1$. These results provide a route toward universal, measurement-free logical control of motional GKP qubits in neutral-atom architectures.
Read the full abstract

All authors Alok Kumar, Aaron N. Raja, Diksha Thapliyal, Ishitwa Kumar Das, Ajay Wasan1

Statistical Mechanics

Effective Hardcore Exclusion Without Exclusion: Two-Species Pair Annihilation Revisited

Reaction-diffusion systems with two species that mutually annihilate through $A+B\to\emptyset$ reactions display a slow decay of the total density $\rho$ which follows an anomalous power law in low dimensions. For hardcore particles in one dimension, this decay follows $\rho \sim t^{-1/4}$ under symmetric diffusion and $\rho \sim t^{-1/3}$ under asymmetric diffusion without relative bias between the two species. The $t^{-1/3}$ behavior, in particular, has so far been observed exclusively in systems with hardcore exclusion. Here we construct a model of particles without hardcore exclusion that reproduces this same $t^{-1/3}$ scaling. In our model, both species undergo asymmetric diffusion with a rate that depends nonlinearly on the local density, a form of transport that, in the absence of the other species, is superdiffusive and falls into the Kardar-Parisi-Zhang universality class. The scaling behavior of $t^{-1/3}$ occurs when at least one of the two species undergoes asymmetric diffusion induced by microscopic processes involving pairs of particles, while asymmetric processes involving groups of three particles lead to $t^{-1/4}$ scaling associated with symmetric diffusion. This demonstrates that the pair annihilation density decay exponent is not determined exclusively by the transport properties of the isolated particle species.
Read the full abstract

Quantum Physics

Gate-Efficient Implementation of the Query-Optimal Time-Dependent Hamiltonian Simulation

The query-optimal algorithm of [CGWZ26] for general time-dependent Hamiltonian simulation uses $$ q = O\left( \alpha T + \frac{\log(1/\varepsilon)}{\log\left(e + \log(1/\varepsilon)/(\alpha T) \right)} \right) $$ queries to $\mathrm{HAM\mbox{-}T}$ within $\varepsilon$ error for a Lipschitz-continuous time-dependent Hamiltonian $H(t)$ on $[0,T]$ satisfying $\left\lVert H(t)\right\rVert\leq\alpha$. However, its direct circuit implementation incurs a substantially larger gate overhead. In this note, we give an implementation of the same algorithm that retains its optimal query complexity and uses $$ O\left[ q \left( a + \log\left(1 + \frac{T(\alpha + \beta T)}{\varepsilon} \right) \right) \right] $$ one- and two-qubit gates, where $a$ is the number of block-encoding ancilla qubits and $\beta$ is the Lipschitz constant of $H$. The main ingredient is an exact dyadic factorization of the ordered update product in the underlying one-query transducer.
Read the full abstract

All authors Boyang Chen, Minbo Gao, Zhengfeng Ji, Tongyang Li, Xinzhao Wang, Shuo Zhou

Mesoscale and Nanoscale Physics

Predictive wavelength tailoring of uniform GaSb-based quantum dots for emission at 1.55 um

A detailed study of emission wavelength tailoring of GaSb-based QDs formed by InGaSb-filling of droplet-etched nanoholes in AlGaSb is presented. The study shows that the emission wavelength can be modified from 1.48 um to the center of the telecom C-band at 1.55 mm by independently varying the QD composition and size. More specifically, the optical transition energy shifts linearly as a function of In-content of the QD material at a rate of -4.4 meV/In-percentage, and with the number of monolayers (ML) of material used for filling the nanoholes, at -2.0 meV/ML. These experimentally observed energy shifts are well predicted by simulations yielding rates of -4.3 meV/In-percentage and -2.1 meV/ML, respectively. For the simulation, a uniform In composition, low intermixing, and microscopically measured QD geometry is considered. Additionally, excellent ensemble QD uniformity, with unprecedented inhomogeneous broadening well-below 7 meV across all samples is demonstrated. Finally, photoluminescence of single-QDs reveals narrow excitonic emission lines of 13.8+/-6.7 ueV and low fine-structure splitting values reaching <10 ueV. These results identify GaSb-based LDE QDs as a tunable telecom platform for scaling quantum-photonic applications over long-haul optical fiber networks.
Read the full abstract

All authors Markus Peil, Maja Wasiluk, Ziemowit Olinkiewicz, Tymon Przychodni, Robert Matysiak, Teemu Taskinen, Joona Salonen, Abhiroop Chellu, Metin Patli, Joonas Hilska, Anna Musiał, Michał Gawełczyk, Mircea Guina, Teemu Hakkarainen

Computation and Language

REER-PT: Reverse-Engineered Reasoning for Perplexity-Guided Pre-training Data Augmentation

As language-model compute continues to scale, high-quality training data is becoming an increasingly important bottleneck. Conventional next-token prediction supervises what follows a context but leaves the intermediate reasoning behind that continuation implicit. We introduce REER-PT, a scalable framework that extends Reverse-Engineered Reasoning (REER) to raw pre-training data. REER-PT identifies continuations that are difficult to predict but can still be inferred from the preceding context, and inserts concise reasoning annotations that reconstruct the missing connection between context and continuation. Candidate annotations are generated and refined offline, with perplexity serving as the optimization signal. Constraints on length and target leakage filter out unhelpful or trivial annotations. This sparse transformation preserves the source text and remains compatible with standard next-token prediction, avoiding online reasoning rollouts during pre-training. We apply REER-PT to transform a source pre-training corpus into an augmented one. Across augmented-data, original-token, and selected-continuation comparisons, perplexity reductions range from 0.42 to 7.29, and only about 0.05% of annotation 13-grams appear verbatim in the source text. We then train two 680M-parameter models with the same architecture and training configuration on the source and augmented corpora, respectively. The augmented-data model gains up to 2.07 percentage points on several knowledge and reasoning benchmarks. Together, the perplexity analysis indicates improved continuation predictability, while the controlled pre-training experiments suggest that this augmentation can improve model performance without changing the standard pre-training objective.
Read the full abstract

All authors Haoran Que, Jiajun Shi, Ting Huang, Renming Pang, Jiaheng Liu, Ge Zhang, Wenhao Huang, Shen Yan, Wei Ye, Shikun Zhang

Applied Physics

A model-inversion control strategy to attain desired equilibrium points when synthesizing nonlinear resonators

The initial conditions of a deterministic nonlinear system affects the attained asymptotic solutions. This is a problem when the nonlinear behaviour is achieved by active control synthesis, and the initial conditions are not accessible for pre-constraining. Meanwhile, the generalized impedance control is recently gaining increasing attention in acoustics, and has been enlarged to achieve nonlinear mechanical responses of Electroacoustic Resonators at low excitation levels. Nevertheless, as the initial conditions of Electroacoustic Resonators are hardly accessible, the equilibrium points actually showcased by the synthetic nonlinear dynamics are not advantageous for noise attenuation in cavities excited by realistic noise sources. In this paper, we propose a control strategy being able to attain the desired equilibrium points, due to the definition of an exosystem characterized by a desired nonlinear impedance and a synthetic external excitation. The exosystem dynamics is integrated in real time and its response is enforced in the actual system (in our case, the Electroacoustic Resonator), by a model-inversion feedforward strategy. The proposed algorithm is implemented in the Electroacoustic Resonator, and is validated both numerically and experimentally. In particular, the experimental testing is conducted both in a quasi-open acoustic environment, and in enclosed cavities for acoustic mode attenuation.
Read the full abstract

All authors Maxime Morell, Emanuele De Bono, Emmanuel Gourdon, Manuel Collet, Alireza Ture Savadkoohi, Claude-Henri Lamarque

Analysis of PDEs

Revisiting Fazly-Wei-Xu pointwise estimates for the fourth-order Hénon equation via Bernstein's technique

In a remarkable work published in Analysis and PDE 8 (2015) 1541-1563, M. Fazly, J. Wei, and X. Xu establish the following pointwise estimate \[-\Delta u \geq \frac 2{n-4} \frac{|\nabla u|^2}u + \sqrt{\frac 2{p+1- \frac 8{n(n-4)}}} |x|^{\frac \sigma 2} u^\frac{p+1}2 \quad \text{in } \mathbf R^n\] for any bounded, positive, $C^4$-solution $u$ to the the fourth-order Hénon equation \[\Delta^2 u = |x|^\sigma u^p \quad \text{in } \mathbf R^n\] with $n \geq 5$, $\sigma \geq 0 $ but implicitly near $0$, and $p>(n+4+2\sigma)/(n-4)$. Their argument relies on a sophisticated iteration argument in the fashion of the standard Moser proof. In this paper, we provide an alternative approach which relies on the Bernstein technique via the maximum principle. The heart of our argument is a suitable choice of an auxiliary function allowing us not only achieving the same pointwise estimate, which still holds even for $p = (n+4+2\sigma)/(n-4)$, but also relaxing the boundedness of solutions. In addition, the pointwise inequality actually holds for any $\sigma \geq 0$. This estimate leads to an interesting consequence of the existence of conformal metric of $\mathbf R^n$ having positive scalar and $Q$ curvatures. We also show that the coefficient $2/(n-4)$ is sharp, hence providing the first answer to the question in the op. cit. paper. Our proof appears to be simpler and constructive.
Read the full abstract

Statistical Mechanics

Time-delayed feedback turns Arrhenius escape logarithmic

Thermal escape is governed by the Arrhenius law, where the mean escape time scales exponentially with the barrier height. We show that the non-Markovianity induced by time-delayed feedback in the confining force removes this exponential scaling. Beyond a threshold set by the curvature of the minimum, the delay destabilizes the well, and the thermal noise seeds an instability that is subsequently amplified deterministically to the boundary leading to slingshot escape trajectories. The escape time becomes logarithmic in the barrier, its fluctuations follow a Gumbel law, and an optimal delay enables escape faster than free diffusion. Our results propose time delay as a tunable and experimentally feasible control parameter for accelerating activated processes.
Read the full abstract

All authors Roy Podgaetsky, Vishwajeet Kumar, Arnab Pal, Ohad Shpielberg

Signal Processing

Compressed Single-Tone Frequency Estimation With Unknown Complex Gain: Singular Rates and Global Identifiability

Fixed linear compression can preserve local Fisher information for a sinusoid yet destroy global frequency identification, while a dark response can make local estimation nonregular. We study both failures for a single complex tone observed through a fixed complex-linear sketch with unknown nonzero complex gain and pre-sketch white Gaussian noise. Near an isolated analytic dark frequency, we separate radial signal vanishing from optimized projective contact between the two signed frequency branches. When the gain magnitude is constrained to a fixed nondegenerate interval, finite contact is equivalent to local quotient identifiability and minimax consistency. The sharp mean-square-error rate is determined by the sum of the radial and contact orders; infinite contact produces exact local aliases. Globally, we formulate a worst-frequency efficientinformation objective on whitened row spaces. We solve it exactly for every even output rank, prove uniqueness and quantitative rigidity of the symmetric-edge projector, and solve the evenaperture co-rank-one case. The rank-two optimum is aliased, and a winding obstruction gives a positive lower bound on the information price of global identification. From three outputs onward, sketches that are globally identifying with an immersive projective response are open and dense and have zero price at the level of suprema; for every even rank of at least four, the exact optimizer has this property. Thus local information preservation, singular recoverability, and global identifiability obey distinct compression laws within one estimation model.
Read the full abstract

Sound

Textual Acoustic Grounding for Generalizable LLM-Based Deepfake Voice Detection

Deepfake voice detection suffers from poor generalization across unseen domains. While Audio Large Language Models (ALLMs) show promise, the modality gap between continuous audio embeddings which capture the subtle acoustic details necessary for deepfake detection and the semantic space of LLMs remains a critical, underexplored bottleneck. We address this by benchmarking diverse audio encoders integrated with Qwen LLMs (0.5B to 7B parameters). First, we demonstrate that fine-tuning the LLM alone risks out-of-domain overfitting, making a frozen LLM a stronger, resource-efficient baseline. Second, to explicitly bridge the modality gap, we introduce a cross-modal prompting strategy that injects linguistic-knowledge-driven acoustic features (via openSMILE) as structured text tokens. This explicit textual grounding not only enhances the frozen baseline but also makes LLM fine-tuning more effective. Ultimately, our approach demonstrates state-of-the-art resilience on the out-of-domain ITW and MLAAD benchmarks, yielding over 16.2% absolute improvement in Macro-F1 over existing ALLM baselines while maintaining competitive in-domain performance. All models reported in this work are \href{https://huggingface.co/01Yassine/AudioLLM-Deepfake-Detection}{publicly available}.
Read the full abstract

All authors Yassine El Kheir, Xin Wang, Wanqing Ge, Tim Polzehl, Sebastian Moeller, Junichi Yamagishi

Computer Vision

Cost-efficient Active Learning for Referring Image Segmentation and Grounding

Collecting natural-language referring expressions along with region annotations, such as masks or boxes, is a major bottleneck in visual grounding (VG), as annotators must write descriptions that distinguish target regions from visually similar ones. We tackle this by formulating active learning (AL) for VG under the realistic setting where only raw images are available without accompanying text. Since ground-truth text is unavailable, sample selection must estimate which images contain ambiguous regions that would require discriminative referring expressions. To address this, we generate auxiliary region-text pairs using foundation models, and introduce Referred Region Ambiguity, a new acquisition function that measures whether the model's confidence collapses onto a single region or disperses across multiple candidates. It allows our method to prioritize images with strong cross-region competition, which are more informative due to their visual ambiguity. We also design a referring-expression annotation interface that helps annotators quickly focus on writing discriminative language with a few clicks. Experiments on RIS and REC benchmarks show that our AL framework consistently outperforms several AL baselines, while a user study shows up to 1.6X faster description labeling of ours.
Read the full abstract

All authors Junbeom Hong, Seonghoon Yu, Hyung Rok Jung, Sundong Kim, Jeany Son

Geometric Topology

On the Nielsen realisation problem for cyclic groups of prime order

We solve the Nielsen realisation problem for high-dimensional aspherical manifolds in a new class of special cases. Mainly, we focus on actions of cyclic groups of prime order. A novelty is that it gives topological solutions to the problem with certain high-dimensional fixed point sets, whereas previous solutions of this type were restricted to discrete fixed point sets. This is enabled by recent joint work with Kirstein on isovariant Poincaré duality spaces, and Farrell-Lück-Steimle's obstruction theory to finding approximate fibrations within a homotopy class.
Read the full abstract

Computation and Language

Hidden Threat in Synthetic Data: Covert Targeted Bias Injection through Benign Text

Synthetic data is increasingly used to train large language models (LLMs), yet its security implications remain poorly understood. Prior work on subliminal learning suggests that models can inherit behavioral traits from seemingly unrelated training data. In this work, we investigate whether such mechanisms can be exploited to inject targeted social biases into aligned models through semantically benign synthetic data. We construct a pipeline in which a misaligned teacher model generates filtered synthetic datasets across domains such as creative writing and code generation, which are then used to fine-tune aligned student models. Our experiments show that benign-looking synthetic data can act as a covert channel for transmitting targeted biases while largely preserving the student model's general task capabilities. These results reveal a previously underexplored security risk in synthetic data-driven LLM training pipelines and highlight the need for improved safeguards. As one possible step toward this goal, we suggest that log-linearity-based scoring may provide a useful signal for screening seemingly benign synthetic data.
Read the full abstract

All authors Minkyung Cho, Jihyo Kim, SeungWoo Song, Junghun Yuk, Minjoon Kee, Hoyun Song, KyungTae Lim

Computer Vision

AQ3D: Adaptive Query Transformer for 3D Instance Segmentation

Transformer-based decoders for 3D instance segmentation typically commit to a fixed number of queries and positional modeling calibrated on the training distribution rather than on the scene at hand. Indoor scans vary widely in spatial extent and object count, so a fixed query set over-initializes small scenes and under-initializes large ones, while learned absolute and relative encodings are bound to the training scenes' extents and can saturate. We present AQ3D, which is designed to handle scenes of various sizes during training and inference. Queries are instantiated at a fixed ratio of the scene's superpoints, forming an overcomplete set whose background rejection is entirely left to the decoder. Positional information is encoded using 3D RoPE over quantized metric coordinates, replacing learned bounded lookup tables of prior decoders. Further, we improve the decoder itself by using attribution-based superpoint pooling, a mask refinement branch, and a cosine classifier for background rejection. Experiments show our method sets a new state-of-the-art on validation and hidden test splits across the datasets ScanNetV2, ScanNet200, and ScanNet++V2 among decoder methods trained without additional data augmentation. Code is available at \href{https://github.com/kenomo/aq3d}{github.com/kenomo/aq3d}.
Read the full abstract

Computer Vision

RealCAD: Towards Real-World Image-to-CAD Reconstruction under Domain Shift and Parameter Bias

Reconstructing editable Computer-Aided Design (CAD) models from images is essential for downstream modification, manufacturing, and design reuse. However, existing image-to-CAD methods are developed predominantly on synthetic renderings and face two coupled obstacles: a substantial appearance domain gap between synthetic and real images, and a previously overlooked parameter bias in widely used CAD data. We show that the local normalization adopted by DeepCAD concentrates several geometric parameters around a few discrete values while encoding substantial information in a single scale factor. Consequently, a model can achieve deceptively high parameter accuracy by exploiting these frequent values rather than inferring geometry from the input image. In this paper, we propose RealCAD, a unified framework that addresses these limitations at the representation, image, and feature levels. At the representation level, we redistribute scale information to the corresponding geometric parameters, producing less concentrated parameter distributions in a shared scale space. At the image level, geometry-constrained translation converts synthetic renderings toward the real-image domain while conditioning on object contours. At the feature level, a multi-positive contrastive objective aligns representations of the same CAD model across viewpoints and image domains, enabling CAD sequence prediction from each individual view. We further introduce OpenRealCAD, comprising four-view photographs of 392 3D-printed objects paired with ground-truth command sequences. Experiments show that the revised representation substantially reduces the accuracy attainable from parameter-frequency priors, making parameter accuracy a more reliable measure of image-conditioned geometric inference. RealCAD further improves real-domain command and parameter accuracy, while retaining competitive synthetic-domain performance.
Read the full abstract

All authors Yihe Sun, Ziyu Lu, Kaihua Tang, Xian-Sheng Hua

Computer Vision

OCR-Based Field Extraction for Archaeological Pottery Metadata: The CENTURIA Dataset

Pottery is a primary source for reconstructing the chronological and economic dimensions of past societies. Archaeologists often document ceramic finds through technical drawings and handwritten metadata. This metadata is critical for dating, provenance attribution, and cross-site comparison, but remains inaccessible to computational analysis, requiring manual transcription of every record. We investigate whether state-of-the-art document analysis models can address this task, and introduce CENTURIA, a dataset of 507 pottery records from the Roman site of Carnuntum, providing transcriptions, bounding boxes, and structured field-level labels across seven metadata categories. Benchmarking five OCR models reveals a substantial domain gap: zero-shot transcription error reaches 15-32% SpACER-M, far exceeding rates on printed archival documents, with domain-specific fields recovered in fewer than 3% of cases. LoRA fine-tuning on just 57 samples, reflecting a realistic archival annotation budget, closes this gap, reducing transcription error to below 1.5% and recovering overall field-level accuracy above 87%. Our results show that a small expert-validated fine-tuning set suffices to convert handwritten pottery documentation into structured, searchable metadata ready for archaeological databases.
Read the full abstract

All authors Gissu Valentina Naghavi, Dominik Hagmann, Martin Kampel, Irene Ballester

Cryptography and Security

Towards Operator-Empowered Vulnerability Hotfixing for 5G Radio Access Networks

Cellular protocol vulnerabilities can remain exploitable for months or years while standards bodies, vendors, and mobile network operators (MNOs) coordinate permanent fixes. We present Buckler, a framework that enables an MNO to deploy temporary, local, and reversible hotfixes in its radio access network (RAN) during this exposure window. Buckler places reusable hooks at standardized L2/L3 channel boundaries and exposes a closed, stateful match-action interface with three preventive actions: DROP, MODIFY, and RELEASE. We evaluate whether this bounded design provides useful coverage without requiring extensive changes to existing RANs. From 23 papers, we identify 64 attacks rooted in standard L2/L3 protocol behavior, of which 43 provide a preventive intervention point at the RAN, and we construct Buckler hotfixes for 20 of them. All 20 hotfixes use the same rule vocabulary and only five standardized channel hooks, while the unsupported attacks expose endpoint dependencies that a RAN cannot satisfy alone. We implement the five hooks on srsRAN and OpenAirInterface with small, structurally similar changes, and demonstrate all three actions against representative availability and privacy attacks. These results establish operator-empowered hotfixing as a practical and portable interim defense and delineate the architectural limits of RAN-only prevention.
Read the full abstract

All authors Dong Hyeok Kim, Xin Zhe Khooi, Hocheol Nam, Seungjin Baek, Mun Choon Chan, CheolJun Park, Min Suk Kang

Computation and Language

TaxCE : A Framework for Automated Taxonomy Construction and Evaluation at Scale

Organizing unstructured feedback text into hierarchical taxonomy is a fundamental challenge in NLP, particularly in domains where feedback arrives at massive scale in varied forms such as reviews, transcripts, and surveys. Existing approaches either produce shallow hierarchies, neglect long-tail topics, or lack rigorous evaluation frameworks. We present TaxCE, a fully automated framework that constructs multi-level hierarchical taxonomies from raw text through progressive condensation of corpus content into actionable segments, deduplicated semantic units, and granular topics with definitions, which are then organized bottom-up into a hierarchy with corpus-groundedness. We also introduce three corpus-grounded evaluation metrics, Exclusivity, Exhaustivity, and Granularity (EEG), and integrate them into a metrics-in-the-loop iterative refinement mechanism that diagnoses deficiencies and applies targeted corrections until convergence. Extensive experiments demonstrate that TaxCE consistently outperforms existing baselines spanning classical topic models, neural methods, and LLM-based approaches, with average improvements of 11.8, 20.5, and 15.7 percentage points in exclusivity, exhaustivity, and granularity respectively over the strongest baseline. Human evaluation further confirms superior taxonomy quality, actionability, and navigability.
Read the full abstract

Digital Libraries

Tracing high-profile attention to questionable research as a case for funder due diligence

When questionable research has entered the scientific record, it stands the same chance as other literature of shaping national and international policy, clinical guidance, and research and development activities. This study uses a known authorship-for-sale network identified in August 2022 to examine whether questionable research influences policy, patents, and clinical guidelines, and whether its authors continue to secure funding and publish after exposure. Across nearly 2,000 publications in our dataset, 57 were cited in policy documents, 12 in clinical guidelines, and 480 in patents. Nearly a quarter of these publications are linked to one or more grants: funding that could otherwise have supported rigorous, ethical science. Among the 278 authors we traced, publishing and funding continued well after the PNN's public exposure in 2022: over 90% continued publishing, and 23% were linked to a grant. As our findings show, the participants of paper mills and authorship-for-sale networks can still inform policy and clinical practice and support their authors' career progression even after the questionable practices behind them are exposed. We present a case for funders to consider authorship and network structure as part of a holistic due diligence process, alongside metrics such as citation counts and attention data.
Read the full abstract

Quantum Physics

Efficient measurement schemes for the Monte Carlo projective quantum eigensolver

The Monte Carlo Projective Quantum Eigensolver (MC-PQE) was recently introduced as an alternative to hybrid algorithms like the variational quantum eigensolver (VQE). By using a quantum Monte Carlo-inspired scheme for energy estimation, MC\nobreakdash-PQE was found to decrease the required measurement cost relative to comparable conventional approaches. However, in the context of VQE, numerous techniques have been developed to reduce the measurement overhead by joint measurement of multiple observables. In this work, we extend these approaches to the asymmetric expectation values required in MC-PQE and assess the performance of different techniques for various molecular systems of up to 12 qubits. We find full commuting Pauli term grouping combined with tailored measurement allocation techniques leads to a 5-10$\times$ reduction in standard error for the same total number of quantum measurements. Conventional Hamiltonian grouped measurement outperforms tractable classical shadows tomography based techniques for the considered systems.
Read the full abstract