Quantum Physics
Lindbladian quantum chaos beyond classical strange attractors
Ángel L. Corps and Armando Relaño
We investigate the connection between classical and
quantum chaos in a dissipative system of two coupled collective
spins. Although the classical dynamics possesses a stable fixed-point attractor, we find pronounced signatures of chaos in the transient dynamics preceding relaxation. In the
quantum model, such transient chaos is accompanied by a transition to random-
matrix statistics in the relevant part of the Liouvillian spectrum and by a qualitative increase in the number of Liouvillian modes contributing to expectation values of physical observables. The transition occurs before the bifurcation at which the stable fixed point disappears and a chaotic attractor emerges. Our results show that Liouvillian
quantum chaos can be associated with transient classical chaos, rather than requiring a chaotic asymptotic attractor, and provide a refinement of the Grobe-Haake-Sommers
conjecture.
Read the full abstract ▾
Quantum Physics
Simplification Rules for Continuous-Time Quantum Walks on Dynamic Graphs
Mostafa Atallah, Daniel Dilley, Jishnu Mahmud et al. (+2 more)
Continuous-time
quantum walks (CTQWs) on dynamic graphs realize
quantum gates as sequences of time-evolving graph
Hamiltonians, but naive constructions produce long sequences with redundancy. Simplification rules, which are graph rewrite rules that shorten a dynamic graph sequence while preserving the
unitary it implements, are the CTQW analogue of
circuit identities in the gate model. In this work, we give CTQW realizations of the standard single-
qubit gates and introduce new graph rewrite rules. We demonstrate the simplification rules through worked
circuit reductions and outline their use as transpilation primitives for converting between the
circuit model and the dynamic graph framework.
Read the full abstract ▾
All authors Mostafa Atallah, Daniel Dilley, Jishnu Mahmud, Zain H Saleem, Rebekah Herrman
Computer Vision
CSWAM: Better Causal Semantic Representations for Out-of-Distribution Generalization in World Action Models
Tianbin Liu, Jian Zhu, Taiyi Su et al. (+5 more)
FastWAM-style world action models enable efficient action-only inference, but generalize poorly under visual distribution shifts. Their reconstruction-oriented representations emphasize appearance-specific details, limiting generalization to unseen scenes and objects. Without observation history, the model also lacks temporal evidence for robustly identifying task-relevant state changes and motion in unfamiliar visual conditions. To address these limitations, we present the Causal Semantic World Action Model (CSWAM), which augments FastWAM with a causal semantic expert built on V-JEPA 2.1. V-JEPA provides temporally grounded representations of semantic state changes and motion with less dependence on appearance-specific details. The expert learns their future
evolution from a sparse history of current and past observations and shares the history-derived context with both the
video and action streams through causal
attention. At inference, CSWAM conditions action denoising on the current
video state and observed semantic history, retaining efficient action-only inference. We conduct simulation and real-
robot experiments to evaluate generalization under distribution shifts. With
embodied pretraining, CSWAM raises Randomized success on RoboTwin 2.0 Clean-to-Randomized transfer from 10.16% to 45.18%, a gain of 35.02 percentage points over FastWAM. Across two real-
robot tasks and three OOD difficulty levels, CSWAM improves average success over FastWAM by 42.5 percentage points, from 27.5% to 70.0%.
Read the full abstract ▾
All authors Tianbin Liu, Jian Zhu, Taiyi Su, Jianjun Zhang, Chong Ma, Zitai Huang, Weiyi Lu, Yi Xu
Artificial Intelligence
Disentangling Long-Term Memory via Latent Neuro-Symbolic Reasoning
Cai Ke, Xinghao Chen, Xiaoyu Shen et al. (+6 more)
Personalized
agents are required to reason over long-term history interactions to infer both explicit preferences and implicit behavioral evidence. While early flat
retrieval methods score
memory fragments independently and neglect the distributed information, current structured
memory frameworks rely on query-agnostic static graphs that fail to capture the context-dependent relations. Crucially, raw textual
memories are inherently entangled and noisy, making fine-grained personalization and cross-session
reasoning computationally prohibitive. To this end, we present LGM, a novel neuro-symbolic framework that shifts long-term
memory disentanglement into a continuous latent space. Specifically, (i) instead of persisting fixed graphs, we design a tailored latent graph construction with a sparse autoencoder. Subject to each query, it maps historical interactions into latent
memory nodes and disentangles the
memory traces into sparse concept activations, dynamically synthesizing query-aware relational edge weights. (ii) A graph encoder then treats the query
embedding as a conditioning preference to direct non-linear message passing across the task-specific latent subgraph. This yields a highly expressive
memory representation for effective activations. Extensive experiments on long-term personalization
benchmarks demonstrate that LGM significantly outperforms state-of-the-art baselines in capturing both explicit and implicit preferences while enabling personalized responses.
Read the full abstract ▾
All authors Cai Ke, Xinghao Chen, Xiaoyu Shen, Keyu Chen, Siyu An, Junnan Dong, Ruifeng Xu, Ruizhi Qiao, Xing Sun
Artificial Intelligence
Collective Loss of Control in LLM Agent Systems: An Epidemic Account of Mutation, Contagion, and Recovery
Xiangfan Wu, Zonghao Ying, Huiyu Wu et al. (+4 more)
How does a multi-
agent system evolve from a local deviation into collective loss of control? We propose an epidemic explanation organized around accidental mutation, contagion, and recovery. A spontaneous deviation creates a seed; communication enables other
agents to adopt and retransmit its unsafe strategy; collective failure can emerge when propagation outpaces correction and containment. Thus, rare individual deviations can coexist with substantial collective risk. Motivated by reported OpenAI
agent coordination incidents, we examine two ingredients of this mechanism. A deployment audit identifies implicit communication paths between nominally independent evaluation runs and verifies transport through a default Docker backend. RogueHandoff-20, a
benchmark of 20 executable scenarios, tests recipient susceptibility by injecting unsafe trajectories generated by a modified Qwen-27B route. Across four native-pending routes, executed harm is 0-5% on normal tasks and 40-95% after injection, exceeding paired direct malicious requests by 5-45 percentage points. These results support low observed baseline harm alongside high conditional susceptibility; they do not establish natural rare-event rates or demonstrate an autonomous cascade. The account motivates complementary defenses: strengthen resistance and recovery alongside prevention of spontaneous deviations, and audit and restrict unintended communication paths that can turn local failures into collective loss of control.
Read the full abstract ▾
All authors Xiangfan Wu, Zonghao Ying, Huiyu Wu, Xing Zheng, Huangsheng Cheng, Xiaorong Shi, Jing Guo
Cryptography and Security
SEEK: Secure and Efficient Encrypted Keyword Search For Privacy-Preserving Messaging Protocols
Soumyadyuti Ghosh and Michail Maniatakos
Encrypted communication protects sensitive user data but can facilitate harmful or unlawful exchanges, creating a trade-off between detecting dangerous messages and preserving end-user
privacy. To address this, we propose SEEK, a practical and efficient encrypted keyword-search protocol for
privacy-preserving messaging that combines homomorphic encryption with secure two-party computation (2PC). SEEK first partitions messages into ciphertext fragments with the minimum sufficient overlap, then homomorphically correlates them using encrypted keyword trapdoors. For long messages, this design can reduce sender-side encryption and upload overhead by up to two orders of magnitude over state-of-the-art baselines. It supports ASCII case-insensitive matching with one fixed-size encrypted trapdoor and one homomorphic multiplication per fragment, yielding up to 5.47x faster correlation computation than the strongest fragmentation-based baselines. SEEK then invokes 2PC-based selected decoding, blinded zero testing, and secure aggregation, revealing only the keyword presence-or-absence bit while hiding the keyword, its length, message contents, match counts, and locations. SEEK achieves 100% accuracy under case variations that result in exact-matching failures, without requiring additional trapdoors or online communication. We further realize SEEK as an end-to-end web and cross-platform mobile application. Prototype evaluation on a weekly messaging history yields an online computation time of 1.92 s per search, demonstrating the practical feasibility and efficiency of SEEK.
Read the full abstract ▾
Combinatorics
A Note on a Result of Chen Li: the Thinning Problem via the Kernel Method
Xuan Fang and Tianyu Wang
Recently, C. Li posted a preprint that resolves the fractional version of Talagrand's discrete convexity problem. In this note, we use the method of C. Li to answer another problem of Talagrand.
Cryptography and Security
AIJon: Automated Generation of Annotations for Fuzzing
Jayakrishna Menon Vadayath, Hulin Wang, Moritz Schloegel et al. (+6 more)
Modern fuzzers use code coverage as feedback to guide their exploration which has proven to be an effective strategy for driving exploration. However, this strategy overlooks inputs that may be interesting to the target program even without uncovering new code paths. Fortunately, prior research has shown that annotations generated by human domain experts can provide additional feedback, guiding the fuzzer towards interesting parts of the program. In this paper, we replicate experiments presented in IJON and extend them to real-world vulnerability detection at scale. To mitigate the scalability challenge, imposed by the need for human domain expertise, we propose utilizing LLMs to automatically generate annotations. We demonstrate the applicability of LLMs for this purpose and observe that LLMs can generate annotations that perform comparably to human-generated annotations. Motivated by this finding, we design AIJON, a system that leverages LLMs to automatically generate IJON-style annotations. We evaluate AIJON on the Magma
benchmark and surprisingly observe that annotation-based fuzzing does not perform strictly better than AFL++. We conduct several experiments to identify the cause of our results and identify key insights regarding the impact of annotations on fuzzing campaigns, including their effect on the energy distribution of the fuzzer. Notably, we observe that LLMs can generate annotations that achieve comparable results to human generated ones, thus opening the door for future research to perform further studies on the impact of annotations at scale.
Read the full abstract ▾
All authors Jayakrishna Menon Vadayath, Hulin Wang, Moritz Schloegel, Jie Hu, Wil Gibbs, Tiffany Bao, Adam Doupé, Ruoyu "Fish" Wang, Yan Shoshitaishvili
Systems and Control
Optimization Design and Simulation Validation of a Variable Stiffness Actuator Based on a Crossed Four-Bar Mechanism
Yuanlong Ji, Ruizhe Jiang, Xiangyu Xie et al. (+3 more)
This paper presents a bio-inspired antagonistic variable stiffness actuator (VSA) based on two crossed four-bar compliant transmission elastic units (CFB-CTEs). The design addresses the difficulty of combining nonlinear elastic shaping with low structural inertia in antagonistic VSA mechanisms. Inspired by the crossed constraint behavior of the anterior and posterior cruciate ligaments during knee flexion, the proposed actuator uses
geometric transmission, elastic energy storage, and bilateral antagonistic arrangement to shape the output torque and equivalent stiffness. A multi-objective optimization model is established to balance torque tracking accuracy, equivalent inertia, and mass. The selected compromise design achieved a torque root-mean-square error (RMSE) of $0.883~\mathrm{N\,mm}$ and a total mass of $50.2~\mathrm{g}$. Its average equivalent inertia was reduced by about 78% compared with an independent torque-only optimized design. An ADAMS multibody model was further built to verify the structural response under single-input, opposite-input, and same-input conditions. The results support the feasibility of the proposed crossed four-bar elastic unit as a lightweight nonlinear elastic branch for antagonistic VSAs.
Read the full abstract ▾
All authors Yuanlong Ji, Ruizhe Jiang, Xiangyu Xie, Junheng Lin, Dongrun Jin, Xingbang Yang
Robotics
ForwardDLO: Model-Based Bimanual Shape Matching of Unconstrained Deformable Linear Objects
Tim Missal, Berk Guler, Lucas Domingues et al. (+3 more)
Ropes, cables, and other deformable linear objects appear in tasks from untangling to cable routing and suturing, yet controlling their shape remains a challenge in
robot manipulation. We study model-based shape control in a general setting: the object lies unfixated on a support surface and two arms may grasp and move it anywhere along its length. Because each arm chooses a grasp point, direction, and magnitude, the joint action space is combinatorially large, and the dynamics model's per-prediction cost bounds how much of it a planner can search. We present ForwardDLO, a recurrent latent dynamics model for this unfixated bimanual setting that predicts per-segment displacements grounded in the observed rope state at every step. Our model reaches accuracy comparable to more expensive baselines while containing no explicit segment-to-segment operations, which makes batched evaluation of candidate actions cheap. On open-loop prediction of real rope motion it reaches the lowest error of the learned models we evaluate, 13% below the strongest baseline. Within a fixed time budget it scores 8 to 22 times more candidate actions than models of comparable accuracy while matching them in real-world shape matching; and on a simulated routing task at a 30Hz control rate, this throughput converts into 98% task success versus at most 30% for the baselines at their own budgets. We release the model, code, and a
dataset of 2.42 million simulated and 14,107 real rope transitions at https://anonymous.4open.science/r/ForwardDLO/
Read the full abstract ▾
All authors Tim Missal, Berk Guler, Lucas Domingues, Simon Manschitz, Jan Peters, Paula Dornhofer Paro Costa
Algebraic Geometry
Virtual classes: An introduction with exercises
Xuanchun Lu and Rachel Webb
We introduce virtual classes in the sense of Behrend-Fantechi, explaining how these directly generalize Fulton's localized top Chern class. We derive Siebert's formula for a virtual class on any Deligne-Mumford stack with a perfect obstruction theory having a global resolution.
Artificial Intelligence
The Mirage of Calibrated Confidence: Trajectory-Independence of Verbalized Confidence in Vision-Language Models
Jisoo Yang, Jaeho Han, Trung X. Pham et al. (+1 more)
A calibrated
Vision-
Language Model (VLM) can repeatedly self-correct, say "Wait, I should recheck," arrive at the wrong answer, and still report high confidence. We find that this occurs because verbalized confidence is largely trajectory-independent in the VLMs and calibration methods we evaluate. We examine this through three complementary lenses: content variation,
token masking, and the model's own hesitation markers. We show that confidence is insufficiently sensitive to what the
reasoning trajectory actually contains, and that calibration training can paradoxically worsen this disconnect. Since existing metrics like ECE and AUROC cannot detect this problem, we propose the Trajectory-Grounding Score (TGS) in two complementary forms: TGS-self, which compares confidence with and without access to the model's own trajectory, and TGS-pair, which tests whether the model assigns higher confidence to correct trajectories than to flawed ones along the
vision,
reasoning, and answer axes. We propose TGS-Bench, a model-agnostic suite spanning 10
benchmarks with controlled good/bad trajectory pairs, and show that conventional calibration rankings diverge from trajectory-grounding rankings, exposing a blind spot in current evaluation practice.
Read the full abstract ▾
All authors Jisoo Yang, Jaeho Han, Trung X. Pham, Junyeong Kim
Number Theory
Extreme Values of Quadratic Dirichlet $L$-functions over Monic Irreducible Polynomials in $\mathbb{F}_q[t]$
Ahammad Mostafa Hossain
In this paper, we establish a new $\Omega$-result for the values $|L(\sigma,\chi_P)|$ of quadratic Dirichlet $L$-functions at and near the central point, where \[ \sigma=\frac12+\frac{\kappa}{\ln g}, \qquad \kappa\geq 0, \] and $P$ ranges over monic irreducible
polynomials associated with hyperelliptic curves of genus $g$ over a fixed finite field $\mathbb{F}_q$. We consider the asymptotic setting in which $q$ is fixed and $g\to\infty$. More precisely, for every $\epsilon \in (0,1/2)$ andevery fixed $\kappa\geq 0$, we prove that \[ \max_{P \in \mathcal{P}_{2g+1}} |L(\sigma,\chi_P)| \gg \exp \left(\left(e^{-\kappa} \ln q \,\sqrt{\frac{\sqrt{q}+1}{\sqrt{q}-1}(1/2-\epsilon)} +o(1)\right)\sqrt{\frac{g\ln_2 g}{\ln g}} \right), \] where $\mathcal{P}_{2g+1}$ is the set of all monic irreducible
polynomials of degree $2g+1$ in $\mathbb{F}_q[t]$. In particular, when $\kappa=0$, this gives the corresponding result at the central point $\sigma=1/2$. Our result extends the recent work of Darbar and Maiti (2024) and yields an improved lower bound for the extreme values in this family. In addition, for $1/2<\sigma<1$ and sufficiently large $n$, we study the extreme values of $L(\sigma,\chi_P)$, where $P\in\mathcal{P}_n$, and obtain an improved lower bound compared with the result of Lumley (2021).
Read the full abstract ▾
Robotics
VLM-MPPI: Grounding Natural Language in Behaviorally Diverse Trajectories for Aerial Navigation
Hanbing Zhang, Fangguo Zhao, Zerui Li et al. (+3 more)
We present a hierarchical
UAV navigation framework that aligns natural-
language intent with dynamically feasible flight behaviors in cluttered indoor environments. To bridge the gap between abstract semantics and low-level control, we employ a parallelized ensemble of six behavior-conditioned Model Predictive Path Integral (MPPI) planners. Crucially, by designing mode-specific guiding costs and sampling biases, we induce distinct trajectory modes that converge to unique behavioral means, yielding a compact set of intentionally diverse candidates rather than mere
stochastic variations. We project these 3D candidates onto the onboard first-person-view RGB stream, turning
language grounding into a visual action selection problem. A pretrained
vision–
language model (VLM) asynchronously selects the candidate index given the overlaid FPV image and a natural-
language prompt, while MPPI replans at 20Hz and a PID-based low-level controller tracks the selected trajectory. We implement the full pipeline in NVIDIA Isaac Sim and on a real-world quadrotor platform equipped with
LiDAR and RGB sensing. Experiments in both simulation and real-world flights show semantically meaningful behavior diversity, robust
language alignment despite VLM latency, and safe, repeatable flight across all modes, achieving 100% task success in our evaluated scenarios.
Read the full abstract ▾
All authors Hanbing Zhang, Fangguo Zhao, Zerui Li, Xin Guan, Peng Cheng, Shuo Li
Fluid Dynamics
Thicker Is Better: How Surfactant Viscosity Governs Uniform Surfactant Lung Delivery
Tristan Beuzelin, Matthieu Labousse and Marcel Filoche
Surfactant Replacement Therapy (SRT) is a well-established intervention for Neonatal Respiratory Distress Syndrome, yet its translation to adults with Acute Respiratory Distress Syndrome (ARDS) has yielded inconsistent results in clinical trials. Using a three-dimensional, zero-dimensional biomechanical model of surfactant bolus transport through the pulmonary airway tree, we demonstrate that, while the tree-like branching architecture of the lung gives rise to discrete flow regimes that result in spatially heterogeneous and inefficient drug delivery, surfactant viscosity is a key determinant of delivery homogeneity: below an analytically derived critical viscosity,
gravity causes complete surfactant exclusion from entire lung subregions regardless of dose volume. A clinically compatible tenfold to thirtyfold increase in viscosity above current formulations drives all bifurcation splitting factors toward the ideal even-split value of 0.5, enabling near-complete and spatially uniform coverage of terminal bronchioles in both infant and adult lung models. Principal component analysis across approximately 24 million simulations confirms that viscosity exerts the strongest positive influence on delivery homogeneity among all injection parameters, with dose volume being the primary determinant of efficiency. These findings provide a mechanistic rationale for reformulating exogenous surfactants with higher viscosity to improve SRT outcomes in adults.
Read the full abstract ▾
Materials Science
A Pseudoscalar Representation Mapping from Parent-Group Vibrational Normal Modes to Symmetry-Adapted Magnetic Structures
Yachao Liu, Haibo Niu and Vei Wang
Conventional approaches classify
symmetry-allowed
magnetic configurations but do not by themselves establish a direct, mode-resolved correspondence with parent-
lattice vibrations. Here, we formulate a universal determinant-induced pseudoscalar twist for all 32 crystallographic point groups, establishing an exact representation-to-geometry correspondence between parent vibrations and
magnetic order. Within the paramagnetic gray group $G\times\Theta_{\mathcal{T}}$, the spatial twist determines
symmetry-defined
magnetic geometry, while time-reversal parity independently specifies
magnetic character. Each parent
phonon irrep $\Gamma$ maps to $\Gamma_{\mathrm{mag}}=\Gamma\otimes\Gamma_{\mathrm{ps}}$, preserving multiplicities and yielding the projection identity $P_{\mathrm{mag},mn}^{(\Gamma\otimes\Gamma_{\mathrm{ps}})}=P_{\mathrm{ph},mn}^{(\Gamma)}$ under the common Cartesian realization. This establishes the Template Principle: parent vibrational modes furnish real-space templates whose
symmetry-enforced nodal
manifolds are inherited exactly. Applied to monolayer $\mathrm{Cd}_2\mathrm{N}_3$, the framework identifies the ferrimagnetic ground state from the parent $A_{2u}$ sector, confirmed by first-principles calculations, alongside cluster
magnetic octupoles and
antiferromagnetic manifolds. It further provides an a priori parent-group criterion for screening
symmetry-allowed linear
magnetic responses.
Read the full abstract ▾
Distributed Computing
From Pixels to Semantics: Edge AI for UAV-Based Critical Infrastructure Inspection
Reza Farahani, Naser Hossein Motlagh, Zoha Azimi et al. (+4 more)
Critical infrastructure assets such as bridges, tunnels, dams, and power line networks require timely and scalable inspection. While conventional manual inspection remains costly and hazardous, unmanned aerial vehicle (
UAV)-based inspection has emerged as an efficient alternative for monitoring difficult-to-access structures. Existing
UAV inspection pipelines have evolved from cloud-centric offline processing toward edge-based perception using lightweight object detectors such as YOLO for real- time defect localization. This article explores the transition toward fully edge-native semantic inspection powered by lightweight
vision language models (VLMs), where
UAVs move beyond object detection toward contextual structural understanding. It categorizes existing
UAV inspection architectures, identifies their key system challenges and architectural requirements, and experimentally assesses the feasibility of semantic edge intelligence on NVIDIA Jetson
UAV-class hardware using the COCO-Bridge
dataset. The evaluation integrates a fine-tuned YOLO-26M for object localization and a lightweight SmolVLM-256 for semantic
reasoning. Finally, it outlines future directions toward agentic, autonomous, trustworthy, and collaborative semantic
UAV inspection across the edge-cloud continuum.
Read the full abstract ▾
All authors Reza Farahani, Naser Hossein Motlagh, Zoha Azimi, Christian Timmerer, Lorenzo Carnevale, Sasu Tarkoma, Schahram Dustdar
Quantum Physics
A Programmable Rydberg Quantum Bus for Nonlocal Connectivity
X. Jin, F. Yang, Weibin Li et al. (+1 more)
Scalable
quantum networks require processing nodes with flexible internal connectivity, yet neutral-atom architectures remain constrained by the strong spatial dependence of native Rydberg interactions. Here we show that a Rydberg atom chain can act as a coherent
quantum bus, converting a locally connected one-dimensional architecture into an effectively nonlocal interaction network. Virtual excitations in the dispersive regime mediate controllable interactions between spatially separated data units, which we derive analytically using a Green's-function continued-fraction method. The resulting mechanism is not restricted to single-excitation dynamics and supports several distinct functionalities, including Floquet-engineered
chiral transport, remote
entanglement of mechanical oscillators, and destructive interference for selectively suppressing unwanted dipole exchange. Simulations incorporating full long-range Rydberg interactions, atomic position fluctuations, and finite Rydberg-state lifetimes show that the mediated dynamics remain robust under experimentally relevant conditions. These results establish Rydberg chains as programmable coherent mediators for extending the internal connectivity of neutral-atom
quantum nodes toward quasi-all-to-all coupling, providing a hardware-level route toward scalable and reconfigurable
quantum-network architectures.
Read the full abstract ▾
All authors X. Jin, F. Yang, Weibin Li, X. Q. Shao
Materials Science
The role of disconnections in redox-induced phase transformation of metal oxides
Martina Ruffino, Baptiste Bienvenu, Xuyang Zhou et al. (+2 more)
Detailed atomic-level understanding of the phase transformation between magnetite (${\rm Fe_3 O_4}$) and haematite (${\rm Fe_2 O_3}$) is lacking, despite widespread interest in redox reactions of metal
oxide systems. While the magnetite-to-haematite transformation entails the addition of oxygen to the system and thus a net diffusive flux, haematite formations have been reported to grow along well-defined habits on the close-packed $\{0001\}_{\rm Fe_2 O_3}//\{111\}_{\rm Fe_3 O_4}$ planes. The propagation of the phase interface in the bulk is thus thought to maintain the oxygen sublattice fixed up to a shear, and to require long-range
diffusion of iron ions. In this letter we propose interfacial steps with dislocation character, i.e. disconnections, as the elementary defects propagating the transformation. We use the topological model of interfacial defects to predict three disconnection modes for the system, and we employ scanning transmission electron microscopy to ascertain their presence in a haematite/magnetite interface of a partially oxidised magnetite powder. Using atomistic simulations, we determine the
equilibrium structures of the disconnections, and study their motion and how they accomplish the transformation, obtaining excellent agreement with atomic-resolution experimental observations.
Read the full abstract ▾
All authors Martina Ruffino, Baptiste Bienvenu, Xuyang Zhou, Dierk Raabe, Yan Ma
Computation and Language
M-SQE: Multilingual Skill Quality Estimation for Enhancing Language Equality in Agentic Skill Use
Yilun Liu, Shimin Tao, Minggui He et al. (+9 more)
Agent skills, reusable procedural documents that extend LLM
agents beyond their parametric
memory, have become an important interface for deploying
agents on real-world tasks. Community-maintained skill libraries built around this interface are growing rapidly. However, this ecosystem remains deeply English-centric: our audit finds that low-resource
languages such as Swahili and Hindi have no in-
language skill content, so
retrieval often returns a skill written in a different
language than the query, degrading accuracy and recall. A practical solution is to synthesize in-
language skills for
retrieval but the quality can be unreliable, so relevance in this setting alone often surfaces a related but unusable candidate. To address this, we propose M-SQE, a post-
retrieval Multilingual Skill Quality Estimation framework that scores candidates via a Theory view for intrinsic quality and an Action view for task-grounded utility, unified into a domain-conditioned final score. We evaluate M-SQE across three skill-use domains: general, tool-use, and cultural tasks. Empirically, we build three-layer candidate skill pools mirroring today's ecosystem, where M-SQE's task success exceeds existing baseline's average by at least +3.5 points across three different retrievers. Particularly, M-SQE lifts the lowest-resource
languages most (+12.9pp on Hindi and +5.6pp on Swahili) and achieves strong performance across all six culture regions, thereby moving agentic skill use toward linguistic and cultural equality.
Read the full abstract ▾
All authors Yilun Liu, Shimin Tao, Minggui He, Chenxin Liu, Li Zhang, Chen Liu, Miao Zhang, Jiaxin Guo, Min Zhang, Liqun Deng, Xiaojun Meng, Daimeng Wei
Computer Vision
DR.WILSS: Diffusion-Based Replay for Weakly Supervised Continual Semantic Segmentation
Leon Arthur Marx, Francesco Barbato, Matteo Caligiuri et al. (+1 more)
Weakly supervised class-incremental semantic
segmentation (WILSS) aims to train a
segmentation model over multiple steps, each introducing new concepts to be learned with only image-level supervision. We introduce DR$.$WILSS, an innovative approach to address catastrophic forgetting in continual learning using
diffusion-based generative replay. Our framework leverages
language clues to guide the
diffusion process, employing self-inpainting and regularization techniques to efficiently produce replay data, aiding the learning process. By generating high-quality replay data, the information from previously learned classes can be preserved during continual updates, a critical challenge in incremental learning scenarios. To further align the statistics of replay data with those of training samples, we apply LoRAs to the generative model. Experimental results demonstrate state-of-the-art performance across multiple
benchmarks and generative architectures, while avoiding storage of training data and the use of additional resource-demanding tools during training. The proposed technique enables an optimal tradeoff between training complexity and inference-time accuracy, making DR$.$WILSS a promising solution for real-world applications.
Read the full abstract ▾
All authors Leon Arthur Marx, Francesco Barbato, Matteo Caligiuri, Pietro Zanuttigh
Instrumentation and Methods for Astrophysics
The GOTO Telescope Control System
Martin J. Dyer, Vik S. Dhillon, Stuart Littlefair et al. (+21 more)
The Gravitational-wave Optical Transient Observer (GOTO) operates a network of robotic, wide-field survey
telescopes searching for optical transients to gravitational-wave signals and other astronomical events. Since 2023, the network has comprised two antipodal sites in the Canary Islands and Australia, each hosting two independent robotic mounts. The GOTO
Telescope Control System (G-TeCS) is a custom Python software package that manages all operational aspects of the network; from alert processing and target scheduling to onsite hardware operations and monitoring. The central system consists of an alert monitor, observation database and scheduler, which issues targets to the four nodes. Each
telescope operates using a system of independent control daemons, with a pilot master control program to coordinate nightly operations. This architecture, in development since 2015, provides a flexible and robust system that now successfully operates the fully autonomous network.
Read the full abstract ▾
All authors Martin J. Dyer, Vik S. Dhillon, Stuart Littlefair, Kendall Ackley, Sergey Belkin, Lisa Kelsey, Tom Killestein, Amit Kumar, Joseph Lyman, David O'Neill, Krzysztof Ulaczyk, Ben Godson, Dan Jarvis, Danny Steeghs, Duncan K. Galloway, Paul O'Brien, Gavin Ramsay, Kanthanakorn Noysena, Rubina Kotak, Rene Breton, Laura Nuttall, Ben Gompertz, Jorge Casares Velázquez, Don Pollacco
Artificial Intelligence
Risk-Aware World Modeling with Flow-Guided Occupancy Evolution for Selective Trajectory Planning in Automated Driving
Rongxiang Zeng, Linsen Cai, Jiafu Zhang et al. (+7 more)
Safe motion planning in automated driving requires anticipating evolving traffic risks and deciding when to revise the current planned trajectory. We introduce RiskWorld, a risk-aware world modeling framework for shared occupancy
forecasting and selective trajectory replacement. Spatial risk fields and temporal actor context are fused with visual bird's-eye-view features. Flow-guided
evolution transports occupancy and scene features, while signed residuals correct occupancy after transport. One forecast is generated per planning step and reused across candidates. Each candidate is compared with a current-state persistence reference, yielding a nonnegative collision-score correction. The trajectory selected by current-world evaluation serves as the planning anchor and is replaced only when additional predicted risk triggers intervention and an alternative satisfies component-wise constraints on predicted risk and trajectory error. Candidate geometries remain unchanged. We evaluate RiskWorld for open-loop planning on nuScenes using camera features, annotation-derived current and historical actor states, and
dataset-provided map context. RiskWorld achieves the lowest collision rate at a long evaluation horizon of 3 s, and the second-best average L2 error among various state-of-the-art baselines, while running at 11.5 FPS on a single NVIDIA RTX 4090 with 90.81 M parameters. Within-setting ablations show that RiskWorld achieves lower collision rates than the current-state rescoring baseline, while forecast reuse enables additional candidates to be evaluated at low marginal computational cost.
Read the full abstract ▾
All authors Rongxiang Zeng, Linsen Cai, Jiafu Zhang, Yijie Zhong, Yide Tao, Shuai Wang, Nan Zheng, Hai L. Vu, Alvaro Garcia Hernandez, Yongqi Dong
General Economics
Multitask Reinforcement Learning for Assisting Choice Model Specification
Gabriel Nova, Stephane Hess and Sander Van Cranenburgh
Discrete choice model specification is a time-consuming task in which modellers often specify and estimate multiple models while balancing goodness-of-fit, parsimony, and behavioural plausibility. We present Delphos, a multitask reinforcement learning framework that learns transferable specification strategies across transport choice
datasets. Delphos frames model specification as a sequential decision-making problem in which it applies a sequence of modelling actions and receives feedback from an estimation environment based on model performance and convergence. To transfer modelling decisions across
datasets with different sets of variables, Delphos represents utility specifications as sets of modelling terms using a DeepSet-Q architecture, allowing a shared specification policy to learn across multiple
datasets. Trained on nine transport choice
datasets, Delphos consistently outperforms independently trained single-task
agents, indicating that sharing modelling experience improves learning efficiency and helps identify promising sequences of modelling decisions with fewer unsuccessful estimation attempts. When applied without further training to the unseen Swissmetro and Decisions
datasets, the same
agent identifies competitive specifications in less than 20 minutes on a standard CPU. It achieves a higher log-likelihood per observation than the VNS metaheuristic on Swissmetro and performance comparable to a published MNL specification developed by expert modellers on Decisions. These findings show that accumulating and reusing modelling experience enables Delphos to function as an intelligent assistant for discrete choice model specification. It reduces manual trial-and-error while allowing modellers to retain control over model diagnosis, refinement, and final selection.
Read the full abstract ▾
Computation and Language
Planning or Improvisation? Stress-Testing the Poetry Planning Site on Open Models and Open Cross-Layer Transcoders
Éric Jacopin
Lindsey et al. (2025) report that Claude 3.5 Haiku plans rhymes: features for candidate rhyme words are active on the newline before a line is written, and a suppress-and-inject intervention redirects the line only when applied there (their Figure 13). We test how far this generalizes on seven
cells crossing four open models (0.6B to 2.6B parameters) with six open cross-layer transcoders (CLTs), on one consumer GPU, decomposing the claim into position specificity (C1), newline site identity (C2), and a newline-resident plan (C3). This is a stress test rather than a faithful reproduction: attribution graphs are unavailable for these CLTs, so features are found bottom-up from decoder vectors. C1 generalizes, in every
cell and in all 247 of 444
prompt-by-inject pairs with a detectable effect, but the effective position is the final
prompt token, adjacent to emission, and only two
cells reach behaviorally meaningful probabilities. C2 and C3 are not recovered by any probe: a census of every active feature finds no rhyme-anticipating enrichment at the newline, and steering the newline while the model composes the whole line, over 36 runs and 8,640 sampled lines, shows why. That intervention is strong but one
token long, making the injected word the first word of the composed line in 703 of 720 samples and leaving the rhyme six words later untouched. A final test drops the transcoder entirely: patching the newline's whole residual, at every layer, from a minimal-pair poem whose third line ends on a different rhyme moves the rhyme in 11 of 1,260 composed lines against 4 at baseline, with a design resolving 1.4%. We read this as a boundary condition rather than a refutation: at this scale and with these transcoders, the causal site is emission-adjacent. We reproduce Figure 13's shape, not its mechanism. Code and data are public (code: github.com/PCfVW/poetry-planning-site).
Read the full abstract ▾
Physics and Society
Topological Uncertainty and Higher-Order Interactions in Spatial Networks
Domenico Pomarico, Alessandro Fania, Gabriel Ramirez Sanchez et al. (+16 more)
Environmental systems are characterized by complex spatial interactions that cannot be fully described through pairwise relationships or local uncertainty measures. We propose a unified framework combining higher-order information theory and topological data analysis to characterize the organization and uncertainty of environmental networks. Spatial entities, represented by monitoring stations or municipalities, are embedded into a Delaunay simplicial complex, and O-information is used to quantify redundancy and synergy among neighboring triplets. The resulting field of higher-order interactions is analyzed through persistent homology, enabling the identification of topological structures that remain stable across interaction scales. The methodology is applied to both an air-quality monitoring network based on weekly \(\mathrm{NO_2}\) and \(\mathrm{O_3}\) observations and a multi-hazard territorial assessment. We show that regions exhibiting strong O-information and persistent topological structures correspond to robust environmental patterns, whereas areas characterized by heterogeneous regimes and rapidly varying interactions display increased uncertainty. Building on these results, we introduce a topological uncertainty framework that integrates simplex divergence, higher-order interactions, and topological uncertainty. Our results demonstrate that uncertainty can be interpreted not only as statistical variability but also as the instability of the underlying information
topology. By integrating O-information and persistent homology within a common spatial framework, the proposed approach provides a new methodology for detecting robust higher-order structures and topologically uncertain regions in environmental and multi-hazard systems.
Read the full abstract ▾
All authors Domenico Pomarico, Alessandro Fania, Gabriel Ramirez Sanchez, Loredana Bellantuono, Domenico Capolongo, Roberto Cilli, Alessandra Costantino, Davide D' Alo, Mario Elia, Francesco Giordano, Niloofar Kheirkhahan, Raffaele Lafortezza, Raffaele Nutricato, Ester Pantaleo, Sabina Tangaro, Roberto Bellotti, Sebastiano Stramaglia, Alfonso Monaco, Nicola Amoroso
Materials Science
Rapid Parameter Estimation from Photoluminescence Decays of Halide Perovskite Thin Films
Robin Heumann, Toby Rudolph, Gaosheng Huang et al. (+2 more)
Extracting material parameters from experimental data is often challenging if no invertible analytical equation can be used to link the data with the quantities of interest. If the link between experiment and material parameters is given mathematically by a set of non-linear differential equations, these must be solved repeatedly during the traditional fitting procedure, resulting in long optimization times and limited insight into parameter uncertainty. Here, we present a parameter estimation workflow specifically aimed at transient photoluminescence measurements performed on lead-halide perovskite films. This workflow is accelerated using artificial neural networks for rapid comparison between experiment and simulation. An advantage of the method is the ability to rapidly scan multidimensional material parameter spaces and identify correlations between parameters that provide insights into the physics of non-radiative recombination in halide perovskites. Finally, we compare steady-state and transient photoluminescence and show how uncertainty in parameters such as the defect density can be reduced by including steady-state data in the parameter estimation workflow.
Read the full abstract ▾
All authors Robin Heumann, Toby Rudolph, Gaosheng Huang, Thomas Kirchartz, Chris Dreessen
Information Retrieval
When RAG Doesn't Help: LLMs for Plausible Material Prediction of Vehicle Components
Frederik Wagner, Sabine Schulte im Walde and Annerose Eichel
In this work, we explore whether LLMs can predict and explain plausible materials for vehicle components such as brake discs or fuel injectors without extensive fine-tuning. We compare a standard generative LLM baseline, single-pass
Retrieval-Augmented Generation (RAG), and an iterative Chain-of-Verification (CoVe) variant across two LLMs, using a domain-filtered Wikipedia corpus that proved helpful in prior work. As no gold standard exists, we develop and release a reusable web-based annotation tool supporting crucial functions for structured domain-expert judgments, such as highlighting, ratings and ranking. Expert agreement is reasonable for plausibility, lower for ratings. LLM-based generation compares favorably to prior work. RAG and CoVe do not improve on it and are partly rated worse by experts. Contrary to prior work, Wikipedia-based
retrieval does not help in our setup.
Read the full abstract ▾
Geometric Topology
Branched real projective structures on surfaces and geometrisation of representations
Gianluca Faraco and Nicholas Rungi
We introduce and study branched real projective structures on compact surfaces. Our main result shows that every representation of the fundamental group of a compact surface into $PSL(3,\mathbb R)$ arises as the holonomy representation of a branched real projective structure, regardless of whether the surface is orientable. We also consider the realisation problem with prescribed branching data, relating the branching degree to the second Stiefel–Whitney class of the representation. The results obtained in this direction suggest several interesting avenues for future research.
Read the full abstract ▾
Artificial Intelligence
WetRobo: A Reproducible Robot Kit for Coding Agents in Biological Laboratories
Yuna Oikawa, Kei Endo, Takanori Uzawa et al. (+5 more)
Automating biological research requires general-purpose, reproducible
robot systems that allow individual wet-lab researchers to delegate
robot tasks without performing teleoperation or neural-network training.
Vision-
language-action policies have been proposed for general-purpose arms, but can lose performance when their operating environment changes. We therefore built WetRobo, a
robot kit that can readily transfer between laboratories. It consists of one
robot arm, laboratory equipment (an incubator, a reagent bottle with a cap, and a Petri dish), the existing code that moves the arm, teleoperation demonstrations of each task that we recorded, and a general
AGENTS.md skill file. A biological experimentalist provides natural-
language tasks without collecting local teleoperation training data or training a neural network. The coding
agent observes the local laboratory and writes and executes programs, using external tools as needed for adaptation. We demonstrate use of WetRobo with OpenAI Codex (gpt-5.6-sol) on three successful tasks: lifting a Petri dish lid, removing a bottle cap, and opening the incubator door, all in real-world laboratories. The coding
agent achieved the cap task in both laboratories, Lab X and Lab Y, whereas a VLA fine-tuned on Lab X demonstrations succeeded there but failed to transfer to Lab Y. These results point to a practical route for laboratory robotics: instead of training a policy for each laboratory, distribute a kit and let a coding
agent adapt it in each laboratory. Code, demonstrations, and the evolved programs are available at https://github.com/tsudalab/WetRobo.
Read the full abstract ▾
All authors Yuna Oikawa, Kei Endo, Takanori Uzawa, Yunzhe Zhang, Manan Anjaria, Lerrel Pinto, Sherry Yang, Koji Tsuda