Peer Revenue Equity in science

Papers

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

All of arXiv

Friday 28 August

Artificial Intelligence

Expert Knowledge & Machine Understanding: Bridging Reactome's Ontology with LLM Semantic Embeddings

Biological knowledgebases like Reactome provide high-quality pathways that include biological elements' relationships and textual descriptions (metadata). The quality of such pathways is granted by manual curation, that presents, however, significant scalability challenges. Lately, numerous NLP tools have been proposed to cope with this issue, leveraging textual information to automatically expand biological knowledgebases. However, little exploration has been done so far to assess whether relationships among textual descriptions mirror higher order biological relationships. This study explores whether human-written descriptions in Reactome can be used to infer the experts' defined global hierarchical structure. To test this, we extracted from Reactome the Homo Sapiens hierarchy of pathways and their reactions (Reactome Hierarchy), and used textual metadata to reconstruct a Semantic Hierarchy, combining a sentence transformer model (SPECTER2) with a modified agglomerative nesting algorithm and a graph reconstruction algorithm. Quantitative (Laplacian Spectral Distance and Bootstrapping) and qualitative (global topological metrics) analyses confirm our hypothesis and indicate that the global hierarchical structure of pathways can be inferred by experts textual metadata.
Read the full abstract

All authors Susanna Bravi, Riccardo De Luca, Rosa Sicilia, Christine Nardini, Mario Santoro

Astrophysics of Galaxies

A High Velocity SiO Jet from S255N SMA3

We present ALMA Band 6 observations of the compact source SMA3 in the S255IR region, analysing both SiO (J = 5-4, 6-5) and CS (J = 5-4) emission to characterise the structure and kinematics of its outflow. The data reveal a highly collimated bipolar jet traced by SiO, reaching line-of-sight velocities of up to $\pm$60 km s$^{-1}$ and de-projected velocities of $\pm$~80 km s$^{-1}$, with a collimation ratio of ~4.7 over a spatial extent of ~5800 au. The inferred dynamical age is 343 yr. Position-velocity diagrams exhibit a continuous velocity structure with triangular morphology, consistent with a steady disk wind rather than an episodic or X-wind-driven flow. Excitation analysis yields a characteristic SiO temperature of ~42 K, and we derive the jet mass, momentum, and kinetic energy which are indicative of a dense, mass-loaded molecular outflow. The emission shows clear chemical and kinematic differentiation: SiO traces the high-velocity jet spine, while CS is observed at lower velocities close to the protostar, and only reaches high velocities at larger distances. Despite forming within a clustered high-mass star-forming region, the jet exhibits a high degree of symmetry, collimation, and dynamical coherence more commonly associated with protostars forming in isolation or less crowded star-forming regions. This demonstrates that well-ordered, narrowly collimated, disk-driven jets can be launched and maintained even in the dense, competitive environments characteristic of high-mass star formation.
Read the full abstract

All authors P. D. Klaassen, M. Reiter, Y. Zhang, K. E. I Tanaka, A. E. Guzman

Rings and Algebras

Orthogonality graphs of real Cayley-Dickson algebras. Part I: Doubly alternative zero divisors and their hexagons

We study zero divisors whose components alternate strongly pairwise and construct oriented hexagons in the zero divisor graph of an arbitrary real Cayley-Dickson algebra. In case of the algebras of the main sequence, the zero divisor graph coincides with the orthogonality graph, and any hexagon can be extended to a double hexagon. We determine the multiplication table of the vertices of a double hexagon. Then we find a sufficient condition for three elements to generate an alternative subalgebra of an arbitrary Cayley-Dickson algebra. Finally, we consider those zero divisors whose components are both standard basis elements up to sign. We classify them and determine necessary and sufficient conditions under which two such elements are orthogonal.
Read the full abstract

Robotics

Picking Bins Empty: A Hierarchical Hybrid Approach with Online Self-Learning of Grasp Points for Reliable Industrial Bin-Picking

Bin-picking is a cornerstone of modern manufacturing, yet achieving complete bin clearance without manual intervention remains a critical challenge. While model-based methods provide high precision, they frequently suffer from deadlocks when predefined grasps are occluded or perception fails. Labor-intensive fine-tuning of grasp points is commonly required to reach a satisfactory performance for new parts. Model-free algorithms offer a more flexible alternative with "out-of-the-box" versatility but lack the reliability and repeatability required for production. Unlike existing work, which treats the two techniques in isolation, we propose a fourtiered hierarchical hybrid approach to combine the best of both worlds. A model-based pipeline serves as a robust backbone, while a model-free "exploration agent" resolves deadlock situations and discovers new grasp points. This is supported by an online self-learning mechanism that uses gripper-stroke feedback and Wilson score intervals to autonomously rank grasp candidates, reducing manual commissioning effort. Validation on three automotive parts demonstrates that our method significantly outperforms a model-free baseline in grasp success rate while improving the bin clearance rate of the model-based baseline from 50.9% to 100% across all experiments. This transition to full bin clearance marks a significant step towards truly autonomous, intervention-free industrial operation.
Read the full abstract

All authors Florian Töper, Samarth Kishor Yelvande, Jan Niklas Ewertz, Rudolph Triebel, Peter Ohlhausen

Computer Vision

Manifold4D: Denoising on Point Cloud Rendered Manifolds for Video Re-shooting

Video re-shooting is the task of re-rendering a monocular video of a dynamic scene along a user-specified camera trajectory. A dominant paradigm explicitly provides the target geometry by lifting the source video into a 4D point cloud using per-frame depth, and rasterizing this representation along the target trajectory to obtain point-cloud renders as geometric conditions for trajectory-controlled generation. The point cloud render and source video may provide conflicting visual cues throughout denoising, as both are directly used as conditions. We refer to this conflict as a trust dilemma, which can compromise trajectory control, 3D consistency of dynamic content, and visual quality, particularly under distribution shifts. We argue that because a point-cloud render is pixel-aligned with the target view, it does not need to be continuously provided as a conditioning signal throughout denoising. We propose Manifold4D, which injects the render directly into the initial noise of flow matching, so that generation no longer departs from standard Gaussian noise but from a geometry-bearing noise distribution that we term the point cloud rendered manifold, leaving the source video as the only visual condition. The render is thus used exactly once, and the network is never asked to learn how to read it; in subsequent denoising steps the model can focus on the source video. In our experiments, Manifold4D attains the best camera-control accuracy on every metric, lowering rotation error by 25–27% and translation error by up to 32% over the strongest baseline, while matching it in video fidelity and attaining the best overall photometric quality and 3D consistency of dynamic content.
Read the full abstract

All authors Yongqi Mao, Zijia Dai, Zhishuo Liu, Wei Xu, Kaiwei Wang, Guotao Meng

Signal Processing

Resource Allocation for Cloud Radar Networks with Communication Constraints

Distributed radar sensing exploits spatial diversity to resolve occlusions and improve estimation accuracy. Realizing these gains, however, relies on the transmission of high-dimensional radar data to a Fusion Center (FC). This imposes significant demands on the wireless network, especially in dense, dynamic, and interference-prone environments like factories, where resilience and latency are critical. This paper studies the resource allocation problem in a capacity-constrained two-hop cloud radar network. We propose a buffered access protocol where sensors perform local spectral windowing to reduce data rates before transmitting measurements to the FC via intermediate Edge Servers (ESs). The resource allocation is formulated as a mixed-integer optimization problem aimed at minimizing the aggregate Cramer-Rao Lower Bound (CRLB) of the target parameters subject to fronthaul and backhaul capacity constraints. We develop an iterative solution algorithm based on Big-M formulation and Successive Convex Approximation (SCA). The proposed framework efficiently identifies the most informative sensor subsets and optimizes time-frequency resource assignments, ensuring high-fidelity sensing within stringent communication budgets.
Read the full abstract

High Energy Physics (Phenomenology)

Three-loop QCD corrections to heavy-to-light form factors and applications to inclusive $B$ decays

We report on the calculation of heavy-to-light form factors at $\mathcal{O}(\alpha_s^3)$ and on selected phenomenological applications in inclusive $B$-decays. After outlining the loop calculation, we extract the hard function in $\bar B \to X_s \gamma$, and discuss our recent progress and preliminary results for the N$^3$LO corrections to partial decay rates in $\bar B \to X_u l \bar \nu_l$, important for the inclusive determination of $|V_{ub}|$. In particular, we establish relations for heavy-quark parameters in the shape-function scheme to four loops and improve a particular model of the $B$ meson shape-function.
Read the full abstract

All authors Matteo Fael, Tobias Huber, Fabian Lange, Jakob Müller, Kay Schönwald, Matthias Steinhauser

Analysis of PDEs

Well-posedness and numerical reconstruction of a source term for linear parabolic problems with an integral constraint

This work investigates a time-dependent source identification problem for linear parabolic equations subject to an integral constraint and Neumann boundary conditions in a domain of $\mathbb{R}^d$, $d\ge 1$. We establish well-posedness and higher regularity of the solution pair in parabolic Hölder spaces. A numerical algorithm based on a finite element discretization in space and an implicit time-stepping scheme is then developed for the reconstruction of the unknown source. The resulting discrete inverse problem involves a well-conditioned operator. We show that identity Tikhonov regularization provides only uniform, nonselective shrinkage in this setting, whereas Tikhonov regularization with derivative-based penalties, analyzed through the generalized singular value decomposition, provides an effective denoising strategy. The regularization parameter is selected using the Morozov discrepancy principle. Numerical errors are evaluated using full parabolic Hölder norms, which provide a more comprehensive assessment of the reconstruction by incorporating errors in the solution, its derivatives, and the associated Hölder seminorms. Numerical experiments for smooth and piecewise constant sources demonstrate accurate and robust reconstructions under increasing levels of noise.
Read the full abstract

Computation and Language

Text Restoration of Ancient Documents with Language Models

Purpose - This study investigates the feasibility of restoring missing text caused by physical lacunae in damaged ancient manuscripts using language models. Methodology - The study proposes different scenarios to replicate real-world conditions. Language models of different architectures are applied according to their suitability to each scenario. We also propose several decoding strategies that further enhance performance and address the discrepancy between lacuna boundaries and the models' tokenization schemes. Findings - The results reveal that text restoration of these documents cannot be fully automated, but it can serve as a useful tool to assist paleographers in their work. Model performance varies greatly depending on which structural part of the document needs to be restored and whether the character length of missing text is available. Originality - This is the first study and to analyze model performance on formulaic and non-formulaic content and the impact of lacuna length awareness in manuscript restoration. Both are recurring challenges in paleographers' manual restoration work. Through systematic comparison and both qualitative and quantitative analysis of different models' performance under varying settings, this study offers a guideline for developing assistive tools to support paleographers.
Read the full abstract

All authors Shibingfeng Zhang, Edoardo Caraffa, Annafelicia Zuffrano, Maddalena Modesti, Giovanni Colavizza

Probability

Characteristic polynomial of self-normalized random matrices

We study the characteristic polynomial of self-normalized random matrices, whose rows are independent and normalized to have unit $\mathrm{L}^2$ norm. The entries before normalization are assumed to have regularly varying tails with tail index $\alpha \in [0,2]$. We prove that, outside the unit disk, the characteristic polynomial converges to a random analytic function $F_\alpha$. We identify $F_\alpha$ as a multiplicative chaos described in terms of Poisson point processes. The family of limiting functions $(F_\alpha)_{\alpha\in[0,2]}$ interpolates between two universal regimes: Poisson multiplicative chaos at $\alpha=0$ and Gaussian multiplicative chaos at the boundary $\alpha=2$. Thus, self-normalization provides a matrix model where one observes the transition between Poissonian and Gaussian regimes for the limiting characteristic polynomial as the tail index varies. A similar transition is found for the fluctuations of the traces of self-normalized matrices. As an application of our results, we derive that the spectral radius of self-normalized matrices is asymptotically bounded above by one in probability for any symmetric entry distribution.
Read the full abstract

Quantum Physics

A quantum generative model for in silico clinical trials using scarce training datasets

In silico methods have emerged as a strategy to complement clinical trials. These are particularly relevant for rare or heterogeneous diseases for which traditional methods are costly or difficult to apply. While classical generative models have shown an extremely good ability to generate high fidelity data when trained using extensive databases, they often struggle when the available samples for training are scarce. In this work, we leverage the potential of quantum computers to represent complex probability distributions to generate high fidelity in silico patients. We propose a pipeline able to combine asymmetric databases into a quantum circuit that serves as a quantum generative model. We evaluate the efficacy of our proposal using a database of Myelodysplastic Syndrome (MDS) patients with 7 clinical variables as a proof-of-concept. We executed our quantum generative model in the IBM Heron r2 “ibm\_basquecountry” superconducting quantum computer and compare our method with well known classical baselines. Our results show that the quantum generative model surpasses the classical generative models in generalization and expressivity metrics, indicating its potential validity to generate high fidelity in silico patients for clinical trials.
Read the full abstract

All authors Olatz Sanz Larrarte, Reza Dastbasteh, Roberto Sanchez-Navarro, Maria Diez-Campelo, Felipe Prosper, Ana Alfonso-Pierola, Mikel Hernaez, Sara Capponi, Pedro Crespo Bofill, Josu Etxezarreta Martinez

Optics

Three-dimensional optical vortices in nonlinear Bessel waveguides

We explore the formation, stability, and propagation dynamics of light bullets carrying orbital angular momentum (vorticity) in Bessel-structured nonlinear graded-index waveguides. While these waves are prone to the collapse in uniform three-dimensional media, we demonstrate that their stabilization is provided by a squared-Bessel index profile in the plane orthogonal to the propagation direction, which may play the role of the waveguiding or antiwaveguiding potential. Different solution families are obtained, revealing a variety of power-vs.-propagation-constant diagrams, characterizing the coexistence of multiple families. We analyze the propagation dynamics in a wide experimentally accessible parameter range. Our results contribute to the understanding of nonlinear light localization in multidimensional settings and suggest new possibilities for using structured light in advanced photonic applications.
Read the full abstract

All authors Carlos F. Sánchez, Humberto Michinel, Boris Malomed, Ángel Paredes, José R. Salgueiro

Superconductivity

Two-dimensional quantum Griffith singularity in three-dimensional ZrN$_x$ superconducting films

We report the experimental observation of two-dimensional (2D) quantum Griffiths singularity (QGS) in $\sim$200-nm-thick epitaxial ZrN$_x$ superconducting films. The films possess a rock-salt structure and are three-dimensional (3D) with respect to superconductivity. For each film with $x \gtrsim 1.30$, the low-temperature magnetoresistance isotherms under fields perpendicular and parallel to the film plane cross over at a broad magnetic field range independently rather than at a single crossing point. Despite the macroscopic 3D nature of the superconductivity, the magnetoresistance isotherms at selected adjacent temperatures follow the theoretical prediction of power-law scaling for 2D superconducting systems, rather than that for 3D systems. The effective critical exponent $z\nu$, obtained by analyzing the magnetoresistance isotherms using the 2D power-law scaling, increases with decreasing temperature and diverges as the quantum phase transition is approached. In addition, the resistivity data near the superconductor-insulator or superconductor-metal transitions obey an activated scaling form that describes the quantum phase transition of 2D superconducting systems governed by an infinite-randomness critical point. The QGS in the ZrN$_x$ films is attributed to quenched disorder induced by intrinsic defects, such as Zr vacancies and N interstitials, which creates spatially inhomogeneous superconducting rare regions. The dynamics of these rare regions, which may exhibit effective 2D characteristics near the quantum critical point, dominate the transport properties of the system near the quantum phase transition. Our results provide compelling evidence for the existence of QGS in 3D superconductors and highlight the crucial role of disorder-induced inhomogeneity in determining the critical behavior of quantum phase transitions.
Read the full abstract

All authors Zi-Yan Han, Li-Min Yu, Yu-Cheng Cong, Yang Yang, Zhi-Xiang Sun, Zhi-Qing Li

Artificial Intelligence

CrabOS: An Operating System for Human-AI Co-inhabitation

AI agents are evolving into long-running computational entities that can invoke tools, maintain memory, and complete complex tasks across applications. In real-world settings, completing a task often requires humans and AI to take turns leading its execution. Such alternation depends on the seamless handoff of the work state of the task between humans and AI. Existing agent systems, however, provide humans and AI with separate work environments. AI agents must therefore rely on additional bridges to continue work: either developers build task-specific interfaces to access the work state, or users manually transfer relevant parts of it through screenshots or textual descriptions. Both approaches make handoffs costly and scale poorly. We propose Human-AI Co-inhabitation, a type of work environment that enables humans and AI to seamlessly take turns continuing work on the same task, and design and implement CrabOS to realize this concept. CrabOS represents the work state as natural-language-readable text objects shared by humans and AI, allowing both to access and manipulate it directly through the same auditable interface without bridges. Case studies show that CrabOS elevates support for complex tasks with alternating human and AI leadership from bridge-dependent application-level solutions to native operating-system capabilities, which provide a new foundation for developing and running AI agents.
Read the full abstract

Quantum Gases

Quantum thermodynamics near the border of a one-dimensional Bose gas

We consider an ultracold Bose gas in a half-open potential well, otherwise confined to a quasi-one-dimensional geometry. The Bogoliubov equations for its elementary excitations are solved in the continuous spectrum to explore the particle and energy content near the edge of the gas, beyond the often applied local density approximation. In particular, gradients in the condensate density enhance density-dominated excitations in the border region. We discuss excess (missing) particles and their energy by comparing to suitable reference solutions for quasi-homogeneous systems. The density profile near the edge shows no Friedel oscillations, but a dipolar feature from the spill-out of thermally excited particles. The calculations are performed in the grand-canonical ensemble and in the thermodynamic limit.
Read the full abstract

Rings and Algebras

Orthogonality graphs of real Cayley-Dickson algebras. Part II: The subgraph on pairs of basis elements

We consider zero divisors of an arbitrary real Cayley-Dickson algebra such that their components are both standard basis elements. We construct inductively the orthogonality graph on these elements. Then we show that, if we restrict our attention to at least $16$-dimensional algebras, two algebras are isomorphic if and only if their graphs are isomorphic. We also provide an algorithm to retrieve the Cayley-Dickson parameters of an algebra from its graph.
Read the full abstract

Signal Processing

Fast Time-Domain MLE for Period Estimation of Pulse Trains

Parameter estimation of periodic pulse trains is a critical task in numerous automated sensing and diagnostic applications. While estimation in the time domain provides superior accuracy in low signal-to-noise ratio environments, its high computational complexity frequently precludes its use in real-time systems. This paper investigates algorithmic optimizations to reduce runtime by leveraging recent advancements in computing architectures. Exploiting the inherent sparsity of the signal via sparse matrix multiplication kernels yields a substantial decrease in inference time. Furthermore, by separating the dense matrix projections into sequential cross-correlation and sparse summation steps, we fundamentally reduce both runtime and memory complexity. These optimizations drastically shrink the memory footprint, making time-domain estimation feasible for large datasets in real-time settings.
Read the full abstract

All authors Sebastian Schertler, Daniel Guger, Stefan Schuster, Stefan Scheiblhofer, Mario Huemer, Alexander Haberl, Johann Reisinger, Oliver Lang

Computer Vision

Empowering Local Agriculture: A Deep Learning-Powered Web System for Identifying Bangladeshi Mango Varieties

Mango variety identification in Bangladesh is challenging because closely related cultivars can have similar visual characteristics and images are often captured under varying real-world conditions. This work presents a deep learning-based web system for automatic identification of Bangladeshi mango varieties. We collected 2,013 high-quality mango images (3024x4032 pixels) from local markets and farms and organized them into nine classes, combining Bari-4 and Bari-7 as a single Bari class. The dataset was divided into training (70%), validation (15%), and test (15%) sets, with image augmentation applied to improve model generalization. Three pretrained CNN architectures, ResNet18, ResNet50, and EfficientNetB0, were fine-tuned under consistent training settings. EfficientNetB0 achieved the best performance, obtaining 98.01% validation accuracy and 97.36% test accuracy, compared with 86.47% and 78.55% test accuracy for ResNet18 and ResNet50, respectively. Class-wise F1-scores for EfficientNetB0 ranged from 0.93 to 0.99, while the Bari class achieved an F1-score of 0.97. The selected EfficientNetB0 model has approximately 4 million parameters, making it suitable for lightweight deployment. We integrated the model into a Streamlit web application that enables users to upload a mango image and receive a predicted variety with class probabilities. The system provides an accessible, practical tool for mango identification and demonstrates the potential of deep learning for supporting agricultural applications in Bangladesh.
Read the full abstract

Hardware Architecture

Gen-TAS: A Generative AI-Aided Hardware-Software Task Allocation Framework for FPGA-GPP Heterogeneous Systems

FPGA-GPP heterogeneous systems combine software flexibility with the performance and energy efficiency of reconfigurable hardware. However, determining which application tasks should execute on the GPP or FPGA requires extensive expertise and design-space exploration, particularly when user objectives vary across latency, communication, resource utilisation, and power. This paper proposes Gen-TAS, a knowledge-grounded LLM framework for user-specific FPGA-GPP task allocation. By combining task-graph analysis with RAG, Gen-TAS grounds LLM reasoning in historical implementation knowledge and generates multiple explainable strategies tailored to the specified objectives. Human-in-the-loop selection and a deterministic backend connect LLM-generated decisions to reproducible FPGA SoC implementations. Experiments on CNN and SDR workloads across multiple LLMs demonstrate stable, requirement-driven allocation. Under latency-oriented objectives, implementations following the selected strategies achieve speedups of up to 2.45$\times$ and 92.53$\times$, respectively, relative to the corresponding all-GPP baselines while other objectives select strategies that trade some acceleration performance for FPGA-GPP communication, resource utilisation, or FPGA power.
Read the full abstract

All authors Mary Kong, Yuqin Zhao, Semih Vazgecen, Cristian Sestito, Themis Prodromakis

Instrumentation and Methods for Astrophysics

The Canadian Galactic Emission Mapper: A New 8-10 GHz Telescope to Map the Polarization of the Northern Sky

The Canadian Galactic Emission Mapper (CGEM) radio telescope is mapping polarized Galactic foregrounds from 8-10 GHz to aid in the search for B-modes in the Cosmic Microwave Background (CMB). Here we describe the design and early on-sky characterization of the CGEM optics. CGEM employs an on-axis, hat feed optical design that is a body of revolution (BOR) and exhibits excellent intrinsic polarization purity. We describe how we have optimized the optics with a novel framework that can directly minimize simulated intensity to polarized intensity ($T\rightarrow P$) leakage in angular power spectrum space. The optimized optics exhibit simulated $T\rightarrow P$ leakage that is orders of magnitude below anticipated $C_\ell^{BB}$ for $r = 10^{-3}$ when scaled from 8-10 GHz to the CMB observing window near 95 GHz. We go on to describe the mechanical design of the optics, including a novel secondary mirror support made from Astroquartz composite that has low loss, low dielectric constant, and preserves the BOR symmetry of the optics. We then showcase the early on-sky performance of the optics with observations of the Sun and satellites, which probe the beam to 40 dB down from the peak. A beam model for the deployed CGEM optics based on electromagnetic simulations is in excellent agreement with the on-sky data.
Read the full abstract

All authors Joshua MacEachern, Mandana Amiri, Charles L. Bennett, Guinevere Berg, Mark Halpern, Gary Hinshaw, Gordon Lacy, Thomas J. Rennie, Shuyu van Kerkwijk, Bruce Veidt, Pedro Villalba-González, Janet Weiland, Don Wiebe, Edward J. Wollack, Parham Zarei

Machine Learning

HARTS: A Prefix-Sharing Training System for Agentic RL with Hybrid-Attention Models

Agentic reinforcement learning (RL) generates irregular rollout trees, but trajectory-by-trajectory training recomputes their shared interaction histories. Existing training-time prefix-reuse systems do not jointly address state recovery, dense execution, and distributed training for hybrid-attention models. We present HARTS (Hybrid-Attention RL over Tree Structures), to our knowledge the first system to demonstrate training acceleration from arbitrary rollout-tree prefix sharing on real hybrid-attention models. A two-level planner constructs compression-oriented microbatches, then jointly refines their membership, cross-microbatch sharing, placement, and execution order. Within each sharing window, gradients from dependent microbatches accumulate before one backward pass through the shared prefix computation. Differentiable boundary-state handoff and bounded replay preserve the per-trajectory chunkwise numerical path, while a linear-time planner attains the global minimum number of sequential linear-attention calls under our packed chunkwise-call model. The runtime uses a common compact layout across attention and dense or mixture-of-experts (MoE) layers, with support for distributed parallelism and activation recomputation. Recovering log-probabilities for each original trajectory position preserves RL objectives; for deterministic top-k routing without token dropping, occurrence counts restore MoE objective weights and expert-load statistics. On SWE-bench, HARTS delivers 4.77-10.45x the update throughput of trajectory-expanded training across two sampling schemes on 7.9B- and 124B-parameter hybrid-attention MoE models using 8-64 GPUs. On the 7.9B model, full-model logit differences are comparable to baseline self-rerun variation. The 124B model shows stable online learning trends through step 150 on SWE-bench and $\tau^3$-Bench.
Read the full abstract

All authors Boyuan Meng, Peihua Bao, Hong Liu, Xiaowei Zhu, Chao Wang, Gen Li, Zhenxuan Pan

Soft Condensed Matter

Proximity to Jamming Governs Acoustic Attenuation in Damped Packings

We use particle-based numerical simulations to address a longstanding question regarding the origins of the linear frequency dependence of attenuation in fluid-saturated granular media. We study both the acoustic modes and wave propagation in damped, disordered particle packings. We calculate the damped vibrational modes of packings as a function of frequency, pressure, and grain-contact dissipation. The spatial structure and dissipation of these modes show a clear transition at a pressure-dependent critical frequency from viscous-like continuum behavior to more localized, scattering modes. We also measure how wavespeed and spatial attenuation rate depend on these same parameters. At the same critical frequency, wave propagation also shifts from coherent motion, where attenuation scales quadratically with frequency and linearly with contact damping, to much more incoherent particle-scale motion, where attenuation scales linearly with frequency and sublinearly with contact damping. All of these features, including the transition frequency, are consistent with a large collection of experimental data, which has not been explained by any framework based on grain-scale physics. We refer to this approach as “Jammed-Network Scattering” (JNS), and propose it as a grain-scale framework for understanding the acoustics of fluid-saturated granular media.
Read the full abstract

All authors Colton Kawamura, Derek R. Olson, Anthony P. Austin, Joshua A. Dijksman, Brian P. Tighe, Abram H. Clark

Software Engineering

From Architecture to Binary: Ensuring Cross-Domain Consistency in Model-Based Airborne Software Development

This paper presents an airborne software development approach for manned and unmanned aerial vehicles aimed at reducing inconsistencies across system, model-based functional, and embedded software domains. In environments influenced by standards such as ARP-4754B and DO-178C, these inconsistencies typically stem from insufficient enforcement across domain boundaries rather than missing process definitions. Building on a previously proposed toolchain centered on a relational interface database, we identify recurring failure modes and propose a repository-centered implementation to address them, tailored to small, resource-constrained teams operating without heavyweight process overhead. Each domain is assigned a primary repository with cross-repository references and dedicated CI pipelines that generate, update, and validate the exchanged artifacts. Automated interface updates, differential change notifications, and consistency checks propagate changes with minimal manual effort and surface inconsistencies before the time-consuming code-generation and compilation steps. An initial implementation in an ongoing experimental project is described, with qualitative feedback from its early use.
Read the full abstract

All authors Nils Schlautmann, Viktor Sinitsyn, Benjamin Engelhard, Florian Holzapfel

Computation and Language

FinExam-10K: When Retrieval Helps Financial Reasoning?

Professional financial examinations require models to combine domain knowledge, calculation, and judgment, yet no benchmark covers the full CFA and FRM structure under one protocol. We introduce FinExam-10K, to our knowledge the largest reported English benchmark for this setting, with 10,198 expert-reannotated questions spanning CFA Levels I-III and FRM Parts I-II. We release 5,110 questions and sequester 5,088 for a quarterly maintained leaderboard. To separate coverage from local answerability, we report a 10,198-item Full-Coverage Track and a 7,625-item Context-Complete Reasoning Track, which is the primary basis for claims about reasoning from the supplied record. Across 17 models, the best accuracy is 85.29% overall. On the frozen Hard band, the best score is 34.68% on the Full-Coverage Track and 54.57% on the 372 context-complete items. All 17 models share 47 context-complete failures. Function-RAG and FunctionGraph-RAG rescue hundreds of errors but also overturn many correct answers, producing little or negative net gain. A gate trained only on public data decides from the question and initial response when FunctionGraph-RAG should run. On the 5,088 held-out items, the gate invokes FunctionGraph-RAG for 7.9% of questions and improves accuracy from 70.83% to 71.23% (p = .0446).
Read the full abstract

All authors Yan Lin, Jingyu Sun, Zhongliang Guo, Qing Li, Zhuohan Xie, Yuxia Wang

Robotics

From Small Talk to Rapport: Exploring Robot Self-Disclosure in Collaborative Tasks

People naturally chat while collaborating and share personal information (i.e., self-disclose) to build rapport and maintain social connections. As robots are increasingly developed to work with people, the effective use of these social behaviors to enhance engagement and support teamwork becomes ever more important. While prior work has shown that robot-initiated small talk can benefit human-robot collaboration, less is known about how best to design such small talk. In this work, we explore how self-disclosure may be designed to support small talk within a human-robot team—especially when the robot is an industrial manipulator that lacks anthropomorphic cues and performs physical work. We first developed an LLM-driven manipulator capable of partaking in small talk, adopting either a low-disclosure or high-disclosure strategy. We then conducted a user study (N = 50) to investigate how self-disclosure in small talk influences human-robot dynamics. Unexpectedly, participants disclosed more in the low-disclosure condition and reported stronger teaming and coordination than those in the high-disclosure condition. This effect was more pronounced among users with prior experience teaming with robots. These results suggest that increasing robot self-disclosure does not necessarily foster rapport, social connection, or reciprocal disclosure; other factors, such as prior HRI experience, should be considered.
Read the full abstract

All authors Kaitlynn Taylor Pineda, Anvii Mishra, Brian Chien, Angela Guo, Toluwani Williams, Ziang Xiao, Chien-Ming Huang

Quantum Physics

Engineering Coherence and Ergotropy through Spatial Arrangement of Environmental Channel in a Two-Qubit Quantum Battery

We investigate how the spatial arrangement of environment channel influences the dynamics of coherence and work extraction in a two-qubit quantum battery with resonant qubits coupled through an XY-exchange interaction. Qubit B is attached to a finite-temperature thermal bath, while RTN noise is applied either to qubit A or to qubit B, depending on whether it is in a separated or co-located configuration, respectively. Two complementary product states are considered to initialise the coherence on different qubits. The dynamics are evaluated using local $l_{1}$-norm coherence, global ergotropy and the global-local ergotropy gap. The results suggest that a stronger exchange interaction enhances the redistribution of coherence without necessarily improving retention of extractable work. For the parameter investigated, the greater global ergotropy is retained in the case of the co-located configuration, especially when the initially coherent qubit is not directly affected by the RTN. However, the configuration provides the largest global-local ergotropy gap, which varies with the exchange interaction, initial state and thermal coupling. Suppression of the gap at large $J$ is typically observed in the case of stronger thermalisation, although enhancement of intermediate-coupling may occur. These results establish that total extractable work, local coherence and collective work advantage quantify the distinct aspects of quantum battery performances. The initial distribution of coherence and the spatial arrangement thus provide complementary mechanisms for regulating locally and collectively extractable work in interacting open quantum batteries.
Read the full abstract

Artificial Intelligence

Under-Mattress Temporal Sensing for Next-Day Agitation Risk Scoring in Dementia Wards

Agitation fluctuates over short time horizons in people living with dementia, yet continuous physiological information for anticipating next-day risk is limited. We assessed whether contactless under-mattress signals from the preceding night inform next-day agitation risk and whether preserving minute-level temporal structure improves performance over conventional nightly summaries. We analyzed 423 patient-nights from 65 subjects in a specialized hospital dementia unit using two under-mattress sensing systems. A unified four-paradigm benchmark compared nightly handcrafted summaries, three-period handcrafted features, full-night sequence modeling, and sliding-window multiple-instance learning. Source-specific preprocessing and five-fold patient-grouped cross-validation were used, with performance estimated from pooled out-of-fold predictions. Evaluation included discrimination, calibration, fixed-threshold metrics, and a comparison of period-signal attribution patterns across two temporal models. Full-night sequence modeling achieved the highest discrimination (AUROC, 0.692; AUPRC, 0.849) and balanced accuracy (0.658). Both minute-level pipelines had higher AUROC than nightly summaries, but differences from three-period handcrafted features were uncertain. Cross-model attribution prioritized activity, heart rate, and respiratory rate during the core overnight period. Calibration remained limited. The preceding night's signals supported modest next-day risk discrimination, with minute-level temporal modeling outperforming nightly summaries. Prospective calibration and external validation are needed before use in individual care decisions. This patient-grouped benchmark identifies contactless overnight sensing as a promising biomedical engineering direction for agitation-risk research in hospitalized dementia cohorts.
Read the full abstract

All authors Zhen Liu, Marta Bono, Robbe Decloedt, Ajda Flisar, Maarten Van Den Bossche, Maarten De Vos

Computation and Language

Nested Byte-Level Vocabularies Are Cheap to Deploy and Expensive to Share: A Pre-Registered Negative Result

A byte-level BPE tokenizer is an ordered list of merge rules, so applying only a prefix yields a vocabulary whose token identifiers are the first rows of the full vocabulary. This prefix nesting allows one language model to operate at several vocabulary sizes, use a control token to indicate the active size, and be deployed at any trained size by slicing its embedding and output head. We pre-registered five claims, including margins, seeds, contrasts, and a stop rule, and trained 30 models with 3.1M- and 10.6M-parameter bodies on 200M tokens each. Slicing is numerically exact: across 76 checks, a sliced model reproduces the restricted full model's logits bit for bit and removes 66% of deployed weights without changing latency. However, the shared model trails a fixed-cap specialist by 3.64% bits per byte at 32k against a 1% margin, and by 2.96% at 8k against a 2% margin. A 2x2 ablation separating the control token from output restriction finds that the token changes performance by +0.07% to +0.13%, with all intervals crossing zero, while output restriction costs +0.47% to +1.19%; the factors are substitutes rather than complements. Multi-cap training nevertheless improves robustness: under typographical noise, the same checkpoint degrades 12.5–15.4 points less in its fine mode and outperforms each fixed-cap specialist at that specialist's vocabulary size. A control with neither cap token nor output restriction is equally robust, attributing this benefit to multi-granularity training rather than conditioning. The per-cap penalty tracks each cap's share of training rows, yielding a falsifiable prediction for future work.
Read the full abstract

Machine Learning

Intrinsic Interaction Geometry Controls the Low-Rank Complexity of Softmax Attention

How much matrix rank is required to preserve every bounded value output of normalized softmax attention? We study the unrestricted maximum-row-\(\ell_1\) approximation rank \(r_\varepsilon(A)\), exactly the least rank achieving uniform error over all bounded vector-valued values. Row softmax exposes the intrinsic interaction \(C=P_m(\log A)P_N\), whereas invertible \(Q/K\) gauges leave \(A\) fixed while changing the Euclidean geometry of a chosen query/key factorization. We replace that coordinate-dependent description by a projective residual \(q(C-T)\) and an attained factor-radius size \(\kappa(T)\). For every rank-\(r\) retained interaction with \(\tau(T)<\varepsilon\), we prove $$ r_\varepsilon(A)\le \min\left\{ N,\; C_r\left( 1+\frac{\kappa(T)} {(\varepsilon-\tau(T))^2} \right)^{r/2} \right\}, $$ with the same unknown dimension constant as the underlying weighted Gibbs-row cover. The profile is gauge invariant, termwise no worse than native retained-subspace bounds at the same declared dimension, and has a worst-case sharp \(r/2\) size exponent at fixed \(r\) and \(\varepsilon\). We then measure \(r_\varepsilon(A)\) directly on learned attention using 9,978 certified brackets across BERT, GPT-2, Qwen2.5, and two ViT checkpoints; where certificates do not close, the optimum remains interval-valued. A pre-specified 2,302-cell held-out study further shows that the historical native-coordinate geometry block contains coarse, mostly head-level information but no detectable incremental information beyond a strong calibrated baseline. The new intrinsic descriptor is not evaluated in that study. Together, the theory and measurements distinguish an operator-intrinsic complexity control from a stronger empirical explanation that the learned-head evidence does not support.
Read the full abstract

High Energy Physics (Phenomenology)

Fixed points of the CKM matrix renormalization group running to all orders in perturbation theory

A proof is given that the six fixed points of the massless 1-loop renormalization group running of the CKM matrix in the Standard Model found previously at 1-loop are fixed points to all orders in perturbation theory. A previous proof required an extra assumption on the running which is eliminated in the proof presented here.
Read the full abstract