Cryptography and Security
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Saturday 7 November
Agentic AI systems extend large language models with tools, memory, retrieval, browser or computer-use interfaces, code execution, external services, and multi-agent coordination. Although these capabilities enable agents to act beyond text generation and perform tasks, they also create attack paths through which adversaries can exploit untrusted content, deceptive metadata, poisoned state, compromised skills, mutable dependencies, or malicious code. Such attacks can manipulate planning and tool use, corrupt memory, abuse code-execution environments, and trigger harmful downstream effects. In this paper, we present a systematic review of malicious attacks targeting agentic AI systems. We screened 186 candidate sources and retained 98. Following independent coding and resolver adjudication, we identified 284 atomic attacks, 452 atomic safeguards, and 660 explicit attack–safeguard relationships. We organize attacks into eight causal-mechanism classes while retaining insertion points, activation conditions, abused capabilities, objectives, impacts, persistence, and evidence as orthogonal dimensions. Our analysis shows that tool, MCP, and action-layer attacks account for the largest share of atomic attacks, while the prompt-injection and action-hijacking class has the broadest paper coverage. We identify distinct class-specific activation patterns and separation between insertion and later activation. Safeguard coverage remains uneven: only 278 of 660 relationships are directly assessed, and strong coverage is uncommon. These findings support lifecycle-aware threat modeling, agent evaluation, tool governance, and defense-in-depth security for agentic AI systems.
Artificial Intelligence
Factor Rank: A Canonical Algebraic Generalization of Tree Decision Diagrams
Tree decision diagrams (TDDs), recently introduced by Capelli et al. (SAT 2026), are representations of Boolean functions. They generalize OBDDs by replacing variable orders with vtrees. The width of a TDD counts the distinct subfunctions across each vtree cut. Counting subfunctions is blind to linear structure. We introduce algebraic tree decision diagrams (ATDDs) by replacing this count with the rank of the cut matrix. The maximum of these ranks over all cuts, the factor rank, is exactly the width of the ATDD. Every TDD can be converted into an ATDD of no greater width. An ATDD is a tree tensor network in canonical form. ATDDs are exponentially more succinct than TDDs, and quasi-polynomially more succinct than structured deterministic DNNFs. Our main result is that this gain in succinctness does not cost canonicity. ATDDs are the first strict canonical form for a rank-based representation over arbitrary vtrees. For each function and vtree there is exactly one ATDD, over the rationals and without any genericity assumption. The canonicity admits a local characterization. Whether a tree tensor network is the ATDD of the Boolean function it computes can be checked in polynomial time from its tensors alone. Equivalence checking becomes a test of syntactic equality. The operation set includes polynomial-time model counting, weighted model counting, and a test whether a tree tensor network represents a Boolean function. Capelli et al. asked for the complexity of changing the vtree of a TDD and suggested local vtree changes. We answer the question. After a rotation of the vtree at one node, the new ATDD differs from the old one in two tensors only. Vtree change runs in output-polynomial time both for TDDs and for ATDDs. We formalize and verify our core results in Lean. We provide a prototype implementation of TDD and ATDD compilers as a first empirical validation.
Read the full abstract
Artificial Intelligence
A Gamified Framework for Conflict-Driven Design: Discovering Strategies for Collaborative Resilience with Multi-Agent Learning
Conflicting stakeholder objectives make it difficult to coordinate investments in urban resilience. We model conflict-driven urban design as a non-zero-sum digital board game in which stakeholder agents allocate scarce resources and balance individual payoffs against shared urban development and resilience objectives. We use multi-agent reinforcement learning (MARL) to train policies and examine their behavior across three settings. On a 4 x 4 board, learned policies reach about 84% of the total return found by a simulated annealing benchmark. On an 8 x 8 board with spatial and resource constraints, trained agents favor projects that improve resilience in self-play. In a 12 x 12 simulation initialized from historical maps of Tatsumi, Tokyo, trained agents achieve a mean total score of 452.1, compared with 357.3 for a random-policy baseline across 10,000 simulation runs. Gains in adaptability, sustainability, flood resistance, and greenery come with reductions in vitality and density. A human-in-the-loop interface supports replay and lets participants take over stakeholder decisions, providing a testbed for exploring these trade-offs in participatory urban planning.
Read the full abstract
Computation and Language
Simple Test-Time Refinement for Plot-to-Code Generation via Visual-Code Diagnostics
Plot-to-code generation has been greatly advanced by recent multimodal large language models (MLLMs). However, existing studies mainly focus on one-pass generation or iterative sampling guided by verifier-based filtering, which does not align with the typical human programming workflow of generate first, then iteratively error-location and fix. This naturally motivates a test-time refinement paradigm for plot-to-code generation. In this work, we present the first study on plot-to-code generation via iterative test-time refinement. We propose a Visual-Code Diagnostics framework that enables refinement through visual feedback from source code. Specifically, we first categorize generation errors into four coarse-grained types, and then train lightweight discriminators to recognize and localize these errors based on the rendered visual outputs and corresponding code. Unlike traditional compiler- or debugger-based verification, our approach performs iterative refinement guided by one diagnostics system, enabling more effective semantic and visual correction during test time. Extensive experiments demonstrate that our method consistently improves generation quality across diverse generators and significantly outperforms strong test-time scaling baselines.
Read the full abstract
Computational Physics
Navier-Stokes characteristic boundary conditions for the Spectral Difference method on simplex elements
Unsteady compressible simulations with high-order methods on unstructured meshes require boundaries that let acoustic, entropy and vorticity waves leave the domain without spurious reflection. For the spectral difference (SD) method, the Navier-Stokes characteristic boundary conditions (NSCBC) have only been formulated on quadrilaterals and hexahedra. Their strong-form formalism relies on a line of flux points normal to the boundary face, which does not exist on triangles and tetrahedra. This paper extends the NSCBC to the SD method on simplex elements, including multi-species reacting gases. The characteristic analysis is performed in the physical frame of each flux point of the boundary face, and the normal derivative of the normal flux is evaluated using the Raviart-Thomas divergence operator. The boundary fluxes are recovered from a linear system local to each boundary element, whose matrix depends on the geometry only. The treatment is compared with the tensor-product formulation at a comparable number of degrees of freedom on a series of test cases covering acoustic, vortical, viscous and reacting flows on triangles and tetrahedra. In all cases, the results are in close agreement with those obtained on quadrilaterals and hexahedra, so that the NSCBC become available for the SD method on simplex meshes, which are well suited to complex geometries.
Read the full abstract
Machine Learning
Learning Biological Hierarchies in Single-Cell Foundation Models
Cell types are organized by taxonomic relationships that define lineages in the so-called cell ontology. However, most existing foundation models typically ignore cell-type lineages encoded in this cell ontology. We introduce a biologically informed foundation model called scOntoFM, which embeds hierarchical cell ontology into representation learning. Using Lowest Common Ancestor (LCA) distances from the ontology graph, our framework pairs efficient offline triplet sampling with hierarchical cell-ontology learning. This approach jointly preserves local neighborhoods and shapes the global embedding geometry to reflect the ontology structure. Across diverse cell- and gene-level benchmarks, the model consistently improves performance, with especially strong gains in zero-shot settings. Our model also balances batch integration with biological structure preservation, providing complementary perspectives to existing embedding evaluation frameworks and supporting scalable biological discovery.
Read the full abstract
Software Engineering
Patch correctness in automated program repair: A five-dimensional evidence framework and cross-benchmark audit from classical repair to coding agents
Tests guide automated program repair and determine which candidate patches are reported as successful. Repository-level coding agents extend this feedback loop by generating tests and revising code after observing execution results. This technical report introduces a five-dimensional evidence profile and reporting protocol to audit how test observations support patch-correctness claims across classical repair systems and coding agents. We assemble an audit dataset of 39 benchmark and evaluation artifacts: 32 primary empirical studies, one institutional technical audit, and six prior taxonomic baselines. We code the 33 primary contributions along five dimensions: test provenance and exposure, behavioral scope, oracle basis and dependence, execution integrity, and label adjudication. The central distinction separates checks visible during repair search from evidence reserved for final assessment. Four worked comparative audits show how changed candidate populations, oracle rules, or execution budgets alter reported detection rates. Additional tests can expose missed faults while rejecting valid alternatives or reducing correct-repair yield under a fixed budget. Repository-level audits separate issue reproduction from candidate-patch discrimination and distinguish automated discrepancy detection from requirements-based adjudication. An evidence matrix links each coded decision to supporting passages and records shared benchmark ancestry. We specify a reporting protocol for repair benchmarks and maintenance pipelines that retains uncertain judgments and records validation cost alongside fault detection and valid-patch retention.
Read the full abstract
Systems and Control
A Nested Heterogeneous Multiscale Framework for Power System Dynamic Simulation
Power system dynamics involve multiple timescales, from electromagnetic transients of the network and inverter-based resources to electromechanical oscillations of synchronous generators and quasi-steady-state evolution of the operating point. Simulating these dynamics with a detailed electromagnetic transient model is computationally prohibitive, whereas existing methods reduce the model to focus on a single timescale and lose accuracy when dynamics of different timescales interact. This paper proposes a nested heterogeneous multiscale framework for highly efficient multi-timescale simulation of a power system represented by one unified detailed model. Rather than varying only the integration step, the framework adaptively shifts the simulation resolution among three nested levels so that the dynamics of interest at each stage are always captured, with the slower dynamics estimated on the fly from short detailed simulations at controlled accuracy. Theoretical analysis proves the stability of long steps during quasi-steady-state evolution and quantifies the attainable speedup. On a modified IEEE 39-bus system with frequency regulation and load restoration, the framework accurately reproduces the dynamics of all timescales and runs 15 to 18 times faster than conventional simulation of the detailed model. The accuracy and speedup of the proposed method are further verified on a 390-bus system, demonstrating its scalability to large power systems.
Read the full abstract
Accelerator Physics
Status of Accelerator Automation in the Community: Classification, Survey Results, and the Role of AI and Machine Learning
The increasing complexity of particle accelerators and the growing demand for operational efficiency have intensified the focus on automation in accelerator control systems. However, there is still a lack of a common language to describe and compare the level of automation across different facilities. This article presents a generic classification scheme for accelerator automation, defining five distinct levels ranging from basic operational assistance to full autonomy. To assess the current state of automation in the accelerator community, a structured survey was conducted during the 14th Workshop on Accelerator Operations (WAO 2025). Responses from major laboratories indicate that more than one third of accelerators currently operate at the lowest level 1 (operational assistance), and no facility has yet reached full automation. However, projections show a clear trend toward higher levels within the next decade. Key drivers include optimization algorithms for autotuning, finite-state machines, and closed-loop feedback systems. This study provides both a conceptual framework and a snapshot of current implementation practices, supporting a shared understanding of accelerator automation across the community.
Read the full abstract
Statistical Mechanics
Ising model on Calandre lattice
In this paper, we proposed to consideration a noval (not yet exploited) planar lattice — Calandre lattice with the Ising model applied to it. In the framework of an exact analytical Kramers–Wannier transfer-matrix approach and using an extension of the combinatorial Vdovichenko–Feynman method, we have obtained exact solution for Helmholtz free energy and consequent expression for the principal eigenvalue of the Kramers–Wannier transfer-matrix. In addition, we derived the exact expression for spontaneous magnetization. The usage of both expressions made it possible to perform the comprehensive studies of thermodynamic and magnetic properties of the Calandre lattice itself and a wide variety of its affiliated lattices. Among them, diversified interrelations between Calandre lattice and already known and even some novel lattices. These lattices have also revealed such effects as magnetic frustrations, multiple phase transitions and magnetic reentrancy.
Read the full abstract
Signal Processing
Modified CNN-Learned Surrogate Covariance With Krylov/Lanczos Subspace for DoA Estimation
Recently, learning-based approaches have been employed to improve the robustness of subspace methods in DoA estimation. A Convolutional Neural Network (CNN) was used as a data-driven framework employing an autoencoder to generate a well-conditioned surrogate covariance optimized end-to-end through a differentiable classical estimator, such as the Root-MUSIC. The complete system is designed to estimate the directions of arrival based on a spectral decomposition using the EVD of this covariance matrix. In this work, a new approach is proposed by formulating the spectral stage of the Root-MUSIC algorithm in terms of the orthogonal complement of the signal subspace, thereby eliminating the need for the noise subspace and noise eigenvalues. Furthermore, the full EVD is replaced by a Lanczos tridiagonalization with reorthogonalization, reducing the overall computational cost.
Read the full abstract
Distributed Computing
Firepool: Brokering Warm GPU Capacity for Agentic and RL Workloads at LinkedIn
AI agents and reinforcement-learning (RL) trainers run thousands of short, isolated GPU jobs: tool calls, verifiers, and rollouts. A job often finishes in seconds, but starting a GPU pod for it takes 20-35 s even when the container image is already on the node, and minutes when it is not. A warm pod removes the wait only if it can serve the next job, and keeping pods warm must neither waste GPUs nor leak state between callers. Firepool, the sandbox and execution broker behind agentic and RL workloads at LinkedIn, meets both requirements. It makes pods interchangeable by building them from a shared base image per hardware profile and binding each job's code, extra dependencies, and cluster topology when the job starts. It keeps them warm cheaply and safely by reusing a pod only for the caller that used it last, staging zero-VRAM standby pods in fractional GPU shares alongside active pods, and capping the GPUs that warm pools may hold via queue-driven budgets. Over a sampled peak period in production, Firepool serves 94.7% of 3,206 pod requests from warm or standby capacity. In paired benchmarks with fixed-duration jobs on production H200 nodes, agent and RL loops finish 13x and 10.8x faster than with a fresh pod per job; against a fresh reuse tag per job, warm reuse cuts p90 start latency from 59.1 s to 3.5 s. When a job cannot reuse the previous pod, standby promotion shortens the handover wait from 51 s to 0.94 s. In local and simulated runs, a keep-alive matched to the loop's pauses keeps the latency benefit while holding about half the capacity of an always-on pool.
Read the full abstract
Artificial Intelligence
Multimodal Models Can See, But Can They Think? A Study of Turkish Visual Reasoning
Vision-Language Models have had massive leaps in recent years. Yet, when tested on high-level cognitive tasks like pattern completion, abstract logic, and relational reasoning, they still fall short. This limitation stems from the models' lack of ability to find out underlying rules and maintain constant intermediate steps, rather than merely deciding on the correct option. To research these shortcomings, this study analyzes visual reasoning performance throughout a multilingual evaluation framework encompassing the BİLSEM, MM-IQ, Visual Puzzles, and PuzzleVQA datasets. We evaluate various data mixing strategies and supervised fine-tuning (SFT) methods. We observed that a correct answer does not always mean the reasoning chain is valid. Because of this, we adopt an LLM-as-a-Judge approach to assess the quality of reasoning beyond final answer accuracy. Our experimental results reveal that baseline models often exhibit more stable reasoning compared to fine-tuned counterparts. Most importantly, we observe that SFT can lead to model collapse, manifesting as infinite text loops or empty outputs. The study concludes that current VLMs still struggle with abstract visual reasoning.
Read the full abstract
Cryptography and Security
The Devil Wears Markdown: Detecting Malicious Intent Hidden in AI Agent Skill Instructions through Hybrid Static, Dynamic, and Agentic Analysis
Agentic AI ecosystems increasingly rely on reusable agent skills that combine executable code, natural-language instructions, configuration artifacts, and external tool integrations. These hybrid artifacts create software supply-chain risks because malicious intent may be distributed across instructions, code, metadata, and runtime behavior, making isolated alerts difficult to interpret. We evaluate three open-source scanners on a held-out test set of 140 skills comprising 80 malicious and 60 benign samples, including Cisco Skill Scanner with and without its optional LLM analyzer. The default Cisco configuration achieves a recall of 0.91 with 33 false positives among the 60 benign skills, while the LLM-enabled configuration using Qwen3:8B achieves a recall of 0.85 with 40 false positives. We then present a framework that combines static findings, safety-constrained sandbox evidence, and LLM-based reasoning to assess malicious intent. The framework tracks the origins of findings and relates evidence across analysis layers. In a within-system evidence-source ablation using the same reasoning backend, adding dynamic evidence to skill-content reasoning increases F1 from 0.70 to 0.85. Combining static and dynamic evidence further increases F1 to 0.89, with five false positives among 60 benign skills. Dynamic evidence provides the largest observed improvement. The Full Pipeline detects 68 of 80 malicious skills, achieving 0.88 accuracy, 0.93 precision, and 0.85 recall. It matches the observed recall of Cisco Skill Scanner with LLM while reducing false positives from 40 to five. This system-level comparison does not isolate model or architecture effects.
Read the full abstract
Nuclear Theory
Spin Asymmetries for Twisted-Neutron Capture by Polarized $^3$He Nuclei
We provide a theoretical analysis of a nuclear reaction caused by twisted neutrons that carry an additional Orbital Angular Momentum (OAM) along the propagation direction. Angular momentum selection rules are derived for the thermal-neutron capture reaction on a spin-polarized $^3$He nucleus. We demonstrate that noticeable novel spin effects due to OAM are limited to the region of the size of de Broglie wavelength around the phase singularity of the beam. Spin asymmetries are predicted for a range of experimental constraints.
Read the full abstract
Numerical Analysis
A high-order structure-preserving ALE–SAV finite element method for moving contact line flows
We develop a structure-preserving arbitrary Lagrangian–Eulerian (ALE) finite element method for incompressible Navier–Stokes flow with gravity and moving contact lines, based on a scalar auxiliary variable (SAV) reformulation. The reformulation expresses the geometric energy quadratically in the auxiliary variable and permits linear treatment of the capillary, wetting and gravitational forces. The method combines high-order isoparametric elements in space with a linearly implicit, second-order Crank–Nicolson discretization in time. Second-order extrapolation supplies the predicted geometry and known coefficients of the linear fluid–SAV system. Mesh motion changes the kinetic-energy inner product and requires a compatible discretization of the time-derivative terms in the ALE formulation. We discretize these terms using an $L^2$-isometric velocity transfer to retain the kinetic-energy balance after the geometry update. We also construct a boundary update that preserves area or volume exactly using only linear solves. The resulting scheme satisfies a discrete SAV energy-dissipation law and preserves area or volume exactly. Numerical experiments in two and three dimensions demonstrate high-order spatial and second-order temporal convergence and confirm SAV energy dissipation and area or volume conservation.
Read the full abstract
Information Theory
Quantum channel bounds for constant-composition codes
We extend the quantum-channel optimization of Alrabiah and Guruswami to $q$-ary constant-composition codes. We study explicit channel and flag constructions that recover selected classical asymptotic upper bounds, including the Ben-Haim–Litsyn, moving-subspace and Horn-channel bounds. We give two equivalent formulations, Gram and flag, and prove stability of the obtained upper bounds under permutation transport and generalized shortening. A hierarchy defined by bounding the output-state ranks within each flag improves strictly between consecutive levels at every composition and every positive distance below the Plotkin threshold. We provide finite matrix parametrizations of these bounds (i.e., the number of variables does not depend on the length of codes under consideration). Numerical optimization rapidly produces improvements over the known asymptotic upper bounds throughout all tested nonbinary parameter ranges. For rational input data, each finite level admits near-optimal feasible rational descriptions of polynomial length. Their feasibility can be checked exactly and their objective values computed to any prescribed accuracy in polynomial time; the associated threshold problem with a prescribed gap belongs to $\mathsf{PromiseNP}$.
Read the full abstract
Signal Processing
Sparse Graph Neural Networks With Spectral Virtual Nodes for MIMO Detection
Graph neural network (GNN) detectors achieve high accuracy in large multiple-input multiple-output (MIMO) systems, but exchange messages between all pairs of transmitted symbols, so their complexity grows quadratically with the number of streams. We first propose a sparse GNN (SGNN) detector that restricts each symbol's messages to its most strongly coupled neighbors, so the message count grows linearly with the system size. However, pruning discards small couplings that jointly shape the weak channel directions, which limit detection accuracy. Even on the dense graph, learned messages capture these directions only partially. We therefore propose a sparse GNN with spectral virtual nodes (SVGNN), which filters the node states through a scaled regularized inverse of the channel Gram matrix to emphasize the weak directions. With two neighbors per symbol, SVGNN reduces the number of multiplications by a factor of 18 relative to the approximate message passing GNN (AMP-GNN) in a 64 x 64 MIMO system. Simulations with QPSK, 16-QAM, and 64-QAM over i.i.d. and correlated channels show that SVGNN matches or outperforms both AMP-GNN and expectation propagation, with gains of up to 3.2 dB, and lowers the error floors of AMP-GNN in several configurations. The virtual nodes improve AMP-GNN even without pruning, indicating that message passing alone, whether sparse or dense, cannot fully capture the weak directions.
Read the full abstract
Robotics
MoE-DAgger: Transition-Aware Router-Entropy Gating for Selective Human Correction in Phase-Structured Manipulation
Selective data acquisition in Dataset Aggregation (DAgger) on robots requires deciding which learner rollouts to correct and when to intervene. In manipulation with ordered phases, router entropy from a phase-supervised soft mixture-of-experts (MoE) is insufficient: the same value can reflect either a normal adjacent-phase handoff or support for nonadjacent or earlier phases. We propose MoE-DAgger, which uses causal router history to distinguish these cases without providing the policy with a phase label, progress variable, or hard phase command during execution. The acquisition rule allows an adjacent transition to settle, but requests correction when high entropy persists outside such a transition, the transition times out, or routing persistently favors an earlier phase. Under matched active human-control budgets, MoE-DAgger achieves 84.0% task success on Green and 78.0% on Red, exceeding phase-unaware filtered entropy by 22 percentage points on Green and 26 points on Red, with no observed collisions in 50 rollouts per task. On 50 shared intervention-free traces per task, 83.3% and 80.0% of its first proposals fall on failed rollouts within predefined failure windows when 15 and 25 traces are selected, compared with 70.0% and 72.0% for temporal action-chunk disagreement and 63.3% and 60.0% for ensemble disagreement. Shadow replay and an active no-veto ablation indicate that suppressing queries at normal adjacent handoffs is a key contributor to the improved allocation.
Read the full abstract
Hardware Architecture
AutoC2V: Synthesis-Guided LLM Training for QoR-Optimized C++-to-Verilog Generation
High-level synthesis (HLS) can translate C/C++ kernels into register-transfer level (RTL) designs, yet generating high-quality hardware still relies heavily on manual pragma tuning and repeated synthesis evaluation. We propose AutoC2V, a C++-to-Verilog generation framework driven by real synthesis feedback that enables large language models (LLMs) to directly generate RTL with improved quality of results (QoR) from synthesizable C++ kernels. We construct training data from real synthesis results, including synthesizable C++ source code, the corresponding Verilog RTL, QoR metrics, and optimization objectives. Using this dataset, we align the model toward synthesis-aware RTL generation through Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO). To reduce the high cost of repeatedly invoking real synthesis during GRPO, we design the QoR Prediction Reward Model (QoR RM), which predicts latency and resource metrics by encoding RTL into long-context semantic representations, thereby providing low-cost and continuous QoR reward signals for policy optimization. Experimental results demonstrate that AutoC2V not only substantially improves RTL functional pass rates, but also achieves superior latency and resource optimization compared to the unfine-tuned base model, representative HLS optimization approaches (HLSPilot, LLM-DSE, and AutoDSE), and advanced general-purpose commercial models (Claude Opus 4.8, GPT-5.5, Gemini 3.1 Pro, DeepSeek v4 Pro, and GLM-5.2). These results indicate that combining real synthesis data with efficient QoR reward modeling can further improve the ability of LLMs to generate high-quality RTL.
Read the full abstract
Geometric Topology
Parallel domains in $SL_3(\mathbb{R})$ and the ideal triangle reflection group
For a given positive triple of flags, we associate an open set, called a parallel domain, in the symmetric space of $SL_3 (\mathbb{R})$ whose boundary is parametrized by three spacelike geodesics of the pseudo-Riemannian symmetric space $SL_3 (\mathbb{R})/GL_2 (\mathbb{R})$. This construction is a higher-rank analog of how three points in the ideal boundary of the hyperbolic plane have an associated ideal triangle. As an application, using the fact that positive triples determine a representation of the ideal triangle reflection group into $SL_3(\mathbb{R})$, we show that these parallel domains are fundamental domains for the group action in the symmetric space. Moreover, using the explicit construction and combinatorial properties of the Farey tessellation, we compute the asymptotics for the critical exponent in the symmetric space and the Hilbert critical exponent as the triple ratio of the triple of flags goes to infinity.
Read the full abstract
Machine Learning
Learning When to Think: Adaptive Internal Computation for Reinforcement Learning
Iterative neural policies usually spend the same computation on every decision, although the benefit of additional processing varies across states. We introduce Adaptive Internal Computation (AIC), an iterative policy that learns when to stop refining its latent representation and act. In AIC, we train action selection and stopping jointly, an expected-depth penalty values computation and intermediate value predictions provide auxiliary supervision. We show that when only some states benefit from further refinement, an ideal state-dependent allocator outperforms every state-independent stopping schedule at the same expected depth. With at most eight refinement steps per decision, AIC executes 3.3 steps on average across twelve benchmark tasks and lowers per-decision latency by 31–44% relative to a fixed eight-step policy on four control and robotics tasks. Its mean reward exceeds the best fixed-depth policy on seven tasks and trails it on four. Ablating adaptive stopping and value supervision separately shows that stopping accounts for most of the reward gain on PickCube and AntMaze, where it also executes more refinement steps and that supervision accounts for most of it on RareEvent.
Read the full abstract
Computer Vision
AmodalBox: Automated 3D Bounding Box Annotation under Partial Visibility
Recent progress in vision foundation models (VFMs) has made automated open-vocabulary 3D bounding box annotation increasingly feasible by jointly leveraging LiDAR point clouds and images. However, partial observations remain a major challenge in sparse driving scenes, where visible LiDAR points often cover only a small fraction of the object's surface and provide insufficient cues for estimating its full spatial extent. As a result, existing methods often produce mislocalized boxes or leave objects unlabeled. To address this limitation, we propose \textbf{\Ours{}}, an automated annotation framework that leverages image-generated 3D assets as amodal geometric priors for partially observed objects. By establishing correspondences between input images and rendered views of 3D assets, \Ours{} robustly aligns canonical amodal geometry with partially observed world-space LiDAR point clouds. To improve temporal consistency and reduce computational cost, \Ours{} generates a single asset per instance, estimates its metric scale once at a single frame and propagates it across frames for dynamic instances, and aggregates multi-frame observations for static instances. 3D object detectors trained on annotations from \Ours{} improve performance over strong baselines by 14.1 mAP on the nuScenes dataset and by 12.45 AP at IoU 0.5 for the vehicle class on the Waymo Open Dataset (WOD). Our project page is available at: \href{https://oliver0922.github.io/AmodalBox/}{https://oliver0922.github.io/AmodalBox/}
Read the full abstract
Analysis of PDEs
Propagation phenomena in spatial periodic bistable reaction-diffusion equations around an obstacle
This paper is concerned with propagation phenomena for a spatially periodic bistable reaction–diffusion equation around an obstacle. It is known that an entire solution emanating from a pulsating front converges to the same front as \(t\rightaroow+\infty\) under complete propagation and an additional technical assumption. We remove the technical assumption and prove that the entire solution coming from a pulsating front converges to the same pulsating front as \(t\rightarrow+\infty\) under the complete propagation.
Read the full abstract
Robotics
Selective Cooperative Perception via Blind Spot Reasoning and Agentic Dialogue with Vision-Language Models
Real-time cooperative perception (CP) in multi-agent traffic environments is highly resource-intensive, which posing scalability challenges for large-scale deployment. Efficiency can be significantly improved by selectively performing CP only in safety-critical scenarios and for key occluding objects. Vision-Language Models (VLMs) are well-suited for complex reasoning in dynamic and uncertain environments, making them a promising tool for this selective CP task. This study proposes a VLM-driven agentic framework for blind spot reasoning and dialogue-based CP. A mixed-model architecture is designed that integrates conventional object detection for stable perception, vision-language reasoning for identifying impactful occlusions, and a physics-based geometric module to align reasoning with real-world visibility constraints. The framework also incorporates a Chain-of-Thought (CoT) prompting mechanism to enhance interpretability and intent-aware decision-making. We collect real-world data from connected vehicles operating in urban traffic and compare the driving-critical blind spot identification and dialogue performance of four state-of-the-art VLMs. Experimental results demonstrate that our method achieves over 80% accuracy in driving-critical blind spot detection and over 90% success in dialogue-driven perception sharing. This work provides a scalable, interpretable, and bandwidth-efficient alternative to conventional CP pipelines, and offers new insights into agentic communication design for future connected traffic systems.
Read the full abstract
Artificial Intelligence
ATHENA: Hypergraph-Based Learning for Toddler Autism Classification from Eye-Tracking
In this work we proposed a architecture designed for the classification of ASD in toddlers using eye-tracking data, addressing the critical need for early and precise detection. The proposed model incorporates FMMix, a feature-level augmentation technique that bolsters the model's learning capacity, and in this work we proposed the ATHENA, which captures intricate feature relationships via an adaptive hypergraph framework. Evaluated on two datasets dataset of gaze coordinates-the model achieves state-of-the-art performance, with AUC scores of 99.80% and 99.86%, respectively. It also exhibits high accuracy of 98.45% and 98.56%, sensitivity 98.01% and 98.10%, and specificity 98.74% and 99.02%, respectively. Demonstrating its superior ability to identify subtle gaze patterns associated with ASD, these results significantly surpass existing methods, highlighting the model's potential for clinical ASD screening. Future efforts will focus on validation across broader populations and integration with multi-modal data to further refine diagnostic accuracy.
Read the full abstract
Quantum Physics
MEAT with source and detector correlations
We prove variants of the marginal-constrained entropy accumulation theorem (MEAT) that enable finite-size security analysis of quantum key distribution with source and detector correlations. For source correlations, our results allow for the state prepared in a given round to depend on the current setting choice and a finite number of preceding choices. For detector correlations, our results require memory effects to persist for a finite number of rounds after a registered click, and all detectors to be switched off during those rounds. These results can be combined, providing, to the best of our knowledge, the first method to jointly treat correlated source and detector imperfections in quantum key distribution. Moreover, we retain all the benefits of the MEAT, such as (but not limited to) support for on-the-fly announcements and choice of adaptive tradeoff functions. The resulting key rates can be computed via a single-round optimization amenable to existing numerical methods. We demonstrate these results on qubit BB84, and exhibit competitive key rates even in the presence of correlated source and detector imperfections.
Read the full abstract
Lattice Field Theory
Flavor decomposition of the nucleon axial and tensor charges from lattice QCD
We determine the nucleon axial and tensor charges using four ensembles of twisted-mass clover-improved fermions with approximately equal physical volumes of about 5.5~fm and four lattice spacings, $ a=0.05$~fm, 0.06~fm, 0.07~fm, and 0.08~fm. The mass of the degenerate up and down, strange and charm quarks are tuned to approximately their physical values. We compute both connected and disconnected contributions to determine the isovector and isoscalar charges and their flavour decomposition. For the nucleon isovector axial charge we find $g_A^{u-d}=1.263(34)(26)$, in agreement with the experimental value. For the isoscalar axial charge we find $g_A^{u+d}=0.458(16)(1)$, while for the corresponding tensor charges we find $g_T^{u-d}=0.951(21)(5)$ and $g_T^{u+d}=0.556(16)(5)$.
Read the full abstract
High Energy Physics (Theory)
Monodromy Defects and Bulk Renormalization Group Flows on the Fuzzy Sphere
A monodromy around a defect specifies a symmetry transformation, but the low energy properties of the defect depend on its interaction with bulk degrees of freedom and also itself. We study this dependence on the fuzzy sphere. Along an anisotropy flow, an $O(3)$ half-twist approaches the $O(2)$ half-twist when its twisted components stay light and disappears when they become massive. At large $N$, untwisted spectators give a positive correction to the defect entropy, fixed by the conformal weight of the free defect. The surviving $O(2)$ monodromy admits a relevant charged perturbation that pins the line. Pinning before or after the easy-plane bulk flow gives matching infrared junction data. A pin of the untwisted component survives only the opposite, easy-axis flow. Walls between opposite pinnings test the stability of the candidate ordered line. They are relevant on the pinned $O(2)$ monodromy lines studied and favoured to be relevant, with a smaller margin, on the ordinary $O(2)$ line, while the ordinary $O(3)$ and $O(4)$ walls lie close to marginality. We separate the wall dimension into a tilt stiffness and an angular shape, comparing them with a Ward identity, a two-loop expansion and a large-$N$ saddle. In the fully twisted large-$N$ model, winding weakens the condensate and lowers the wall dimension and the entropy released by pinning. The resulting entropy comparison permits disordering into the unpinned line wherever the leading wall estimate is relevant. Finally, we construct the regular half-twist on the fuzzy four-sphere as a starting point for analogous questions in five dimensions.
Read the full abstract
Signal Processing
A Bussgang-based Latent Lifting Method: Discrete RIS Channel Estimation and Beamforming
This paper addresses the challenge of discrete phase-shift optimization in reconfigurable intelligent surface (RIS)-aided multiuser communication systems. Practical RIS hardware often utilizes low-resolution phase shifters, leading to challenging combinatorial optimization problems. To address this challenge, we propose a Bussgang-based latent lifting (BuLL) method that represents discrete phase shifts as quantized outputs of continuous latent variables. Our BuLL method calibrates the Bussgang gain and distortion covariance under a reference distribution and keeps these moments fixed during deterministic latent optimization. These fixed moments are then used to construct a distortion-aware continuous surrogate that incorporates reference-averaged discretization effects without penalty terms. For uplink channel estimation, we develop a matrix quadratic transform-based algorithm to optimize a mean-square-error surrogate. For downlink beamforming, we develop a generalized power iteration-based RIS phase-shift optimization algorithm for a spectral efficiency (SE) surrogate. Numerical results demonstrate improvements in channel estimation accuracy and sum SE over conventional benchmarks in the evaluated low-resolution RIS settings.