Numerical Analysis
Numerical Realization of an Entropy-Based Selection Principle in the Example of Kelvin-Helmholtz Instability
Shaoshuai Chu and Michael Herty
In this paper, we extend numerical schemes based on parameterized Young measures and linear programming to two space dimensions and apply them to the Kelvin-Helmholtz instability governed by the two-dimensional Euler equations of gas dynamics. We construct first-, second-, third-, fifth-, seventh-, and ninth-order Young-measure schemes and compare them with corresponding standard local Lax-Friedrichs (LLF) flux-splitting schemes and LLF schemes equipped with an
entropy-based local characteristic decomposition (LLF-ELCD). We examine instantaneous and time-averaged density profiles, cumulative averages across spatial reconstruction orders, regional empirical density distributions, averaged density marginals, and several selection criteria for dissipative weak solutions. The numerical results reveal scheme-dependent flow patterns, particularly in the small-scale structures generated during the roll-up of the shear layers. At every reconstruction order considered, the Young-measure schemes yield the largest time-averaged physical
entropy. By contrast, the LLF-ELCD schemes do not systematically yield larger
entropy than the standard LLF schemes and therefore do not constitute a consistent maximum-
entropy selection mechanism. The averaged density marginals of the Young measures are concentrated on several neighboring density states, and the later cumulative averages are closer for all Young-measure schemes. These results demonstrate the feasibility of the two-dimensional Young-measure formulation and show that the objective function in the linear-programming problem can act as an effective selection mechanism. The observed
entropy preference is consistent with the local optimization of the expected physical
entropy. It cannot be reproduced by merely incorporating
entropy into a conventional numerical construction.
Read the full abstract ▾
Numerical Analysis
The transposition method in saddle-point problems
Dietmar Gallistl and Zhen Liu
The transposition method known from the theory of second-order
elliptic partial differential equations is formulated for saddle-point problems under certain structural assumptions known from the Brezzi splitting plus mapping properties that reflect
elliptic regularity in concrete instances of differential operators with boundary conditions. A framework is devised that grants approximation by standard, unmodified Galerkin-type methods, which may be nonconforming with respect to the more general data allowed by the transposed operator. That the data need not be regularized is an advantage particular to the saddle-point approach. Examples covered by the framework include the
Laplacian, the Stokes operator, linear elasticity, the bi-
Laplacian, and the stationary Maxwell problem with $L^2$ boundary data.
Read the full abstract ▾
Human-Computer Interaction
Beyond Screen Time: Demonstrating the Value of App Activity Logs to Understand User Behavior Further
Ole Schmitt, Pauline Gieseler, Frederik Riedel et al. (+4 more)
Research on smartphone use remains fragmented, with studies differing in measures, methods, and platforms, limiting what we know about everyday behavior. We argue that event-level app activity logs should form the backbone of research on actual rather than recalled smartphone use. We conduct a secondary analysis of 4,571,252 app events from 1,972 participants across three longitudinal cohorts differing in age, country, recruitment, and platform, complemented by a published reference cohort. We reproduce established aggregate and micro-usage measures and examine temporal structure, application composition, transitions, and individual distinctiveness. Aggregate usage varies less than the organization of activity: adolescent use, for example, is structured around school schedules, while other cohorts show weaker within-day patterns. Application and transition patterns reveal behavioral differences obscured by screen time; an intervention reduced daily usage while increasing mean session duration. Activity traces also enabled 15-22% top-1 participant re-identification. We discuss methodological, reproducibility, and
privacy implications for HCI.
Read the full abstract ▾
All authors Ole Schmitt, Pauline Gieseler, Frederik Riedel, Donatus Wolf, Ariel Stern, Paul Schmiedmayer, David Grüning
Multiagent Systems
Mixed-integer flow formulations for motion planning and decision-making of networked multi-agent systems
Angelo Caregnato-Neto, Paul-Louis Delacour, Raf Van de Plas et al. (+2 more)
This work investigates the use of flow-based connectivity maintenance constraints in mixed-integer linear programming (MILP) trajectory planning and decision-making models for networked multi-
agent systems (MAS). We integrate flow-based encodings for standard and k-hop connectivity into MILP multi-vehicle maneuvering models that are widely used alongside receding horizon planning strategies. Their necessity and sufficiency is demonstrated, guaranteeing full coverage of potential network
topologies. The flow formulation for standard connectivity decreases the growth of the required inequality constraints from exponential to
polynomial w.r.t. the size of the MAS when compared to the state-of-the-art subtour elimination (SEC) method. The flow-based k-hop connectivity constraints decrease the number of required binary variables and decouple its growth from the number of hops. However, the impact of these formulations in performance is not straightforward due to the introduction of a substantial number of continuous flow optimization variables and, in the case of k-hop connectivity, additional inequality constraints. We investigate this trade-off through a statistical evaluation of costs and optimization times using a conventional branch-and-bound commercial solver and trials performed with randomized environments for increasingly larger MAS. The results show that the flow formulation outperforms SEC in standard connectivity problems, enabling the solutions to be computed for larger MAS considering the imposed optimization time limit. The reduction in number of binary variables enabled by the k-hop flow formulations decreases the theoretical worst-case number of iterations required by the branch-and-bound
algorithm to compute the global optimal solution. Our results show that this advantage did not translate into improvements in the average performance when compared to the baseline.
Read the full abstract ▾
All authors Angelo Caregnato-Neto, Paul-Louis Delacour, Raf Van de Plas, Tamás Keviczky, Janito Vaqueiro Ferreira
Digital Libraries
The illusion of neutrality in metric-based research evaluation
Fengyuan Liu, Hazem Ibrahim, Yasir Zaki et al. (+1 more)
Venue prestige and citation counts are two widely used, albeit imperfect, signals of research quality. When the two signals conflict, evaluators must decide how much weight to assign each. Yet, it remains unknown how researchers across disciplines trade off these two signals when evaluating research outcome. To fill this gap, we surveyed 869 researchers using paired choices between hypothetical departmental hiring rules that assigned different weights to venue prestige and citation counts, asking which would produce better science. Among 795 respondents whose choices were largely internally consistent, choices placed nearly equal aggregate weight on the two signals. This apparent balance concealed substantial individual heterogeneity: nearly two in five respondents occupied the most venue-heavy or citation-heavy intervals. Moreover, respondents on average chose more citation-heavy rules than they believed their departments used in hiring. By revealing the subjective judgments that arise when research indicators conflict, our findings reinforce calls for greater caution when quantitative indicators are used to evaluate research.
Read the full abstract ▾
All authors Fengyuan Liu, Hazem Ibrahim, Yasir Zaki, Talal Rahwan
Computer Vision
Spatial Action Review: A Visual Analytics Dashboard for Auditing Language-to-Action Hand-offs in Electron Microscopy
Samia Mohinta and Albert Cardona
Multimodal large
language models (MLLMs) are increasingly explored as interfaces for scientific image analysis, where a visual question-answering (VQA) response may be paired with a spatial output that guides a downstream stage. A supervisor reads the
language answer, while a downstream workflow such as
segmentation or region review consumes the point-set output. We call this transition from inspecting the answer to relying on its point action the
language-to-action hand-off. A silent failure occurs when the answer is correct while the paired action misses annotated objects needed downstream, so answer-based oversight clears a region whose action is unreliable. We introduce Spatial Action Review, a visual analytics dashboard for
auditing this failure mode in electron microscopy (EM) mitochondria analysis. It links paired answer-action records through an answer-action ledger, a task-by-
dataset risk map, and an image-region audit view, connecting aggregate patterns to image evidence while an adjustable action-reliability gate supports re-audit. The review ends in a human-AI hand-off, where a supervisor records whether the action is accepted, escalated, held under a stricter gate, or flagged for model revision. Across 541 image regions from an EM-adapted Qwen3-VL case-study run, point actions fail the gate in 54.4% of records with a correct VQA response, and 27.4% of all records are silent failures. A correct answer is associated with only a 5.8-percentage-point higher probability of a reliable action, with a bootstrap interval spanning zero; the point-biserial correlation between answer correctness and object coverage is 0.061. This weak coupling persists across five model conditions on 753 matched image regions. Spatial Action Review makes answer-action mismatches visible and ties them to image evidence and a recorded decision before MLLM outputs enter autonomous scientific workflows.
Read the full abstract ▾
Computer Vision
SGAM: Shared Gaussian Geometry with Implicit Amplitude Modeling for Scan-Specific 3D Multi-Contrast MRI Reconstruction
Jingran Xu, Dong Liang, Hairong Zheng et al. (+2 more)
Three-dimensional (3D) multi-contrast
magnetic resonance imaging (MCMRI) provides rich anatomical and quantitative information but requires long acquisition times, motivating k-space undersampling. However, reconstruction of large volumetric
datasets imposes substantial computational and
memory demands. To address this challenge, we propose SGAM, a
memory-efficient, scan-specific framework for joint full-volume 3D MCMRI reconstruction. SGAM is based on a shared-geometry Gaussian representation in which amplitudes are modeled by a multi-output implicit neural representation (INR) and contrast-specific phases by explicit variables. This design exploits common anatomical structure across contrasts while preserving contrast-specific signal variations. The representation is jointly optimized using only the acquired multi-coil k-space without external training data. Experiments showed that SGAM consistently outperformed the comparison methods across imaging tasks and acceleration factors, with greater improvements under stronger undersampling. SGAM also achieved a favorable balance between reconstruction quality and computational cost, demonstrating its effectiveness for 3D multi-contrast MRI reconstruction.
Read the full abstract ▾
All authors Jingran Xu, Dong Liang, Hairong Zheng, Yuanyuan Liu, Yanjie Zhu
Machine Learning
A Temporal Knowledge Graph for Music Festival Lineup Forecasting
Julia Gastinger, Thilo Dieing, Christian Meilicke et al. (+1 more)
Music festival lineups emerge from complex relationships among artists, genres, releases, labels, and past performances, making the prediction of future lineups a natural fit for temporal knowledge graph (TKG)
forecasting. In this work, we present a TKG covering 380 festivals over 55 years, comprising more than 90K festival performance quadruples along with information on festivals, artist tours, and artist metadata, and release it as a resource for TKG
forecasting evaluation. We formalize festival lineup
forecasting as temporal link prediction between artists and festivals at future timestamps. We evaluate six TKG
forecasting models on this task, analyze their capabilities and limitations, and compare them against Large
Language Models applied zero-shot. Our resource complements existing TKG
benchmarks by grounding evaluation in a concrete, real-world application domain.
Read the full abstract ▾
All authors Julia Gastinger, Thilo Dieing, Christian Meilicke, Heiner Stuckenschmidt
Human-Computer Interaction
A Leader-Driven Open Collaboration Platform for Exploring New Domains
Michael Weiss, Ibrahim AbuAlhaol and Mohamed Amin
This paper describes the design and initial evaluation of a leader-driven open collaboration platform for exploring new domains. The goal of this platform is to enable the collaboration of subject matter experts across knowledge boundaries. Traditionally, new domains are explored from within a single specialist or a focused group perspective. However, this often introduces bias. Collaboration helps reduce such bias by providing access to a broader range of information sources, increasing the chances for producing new insights in a new domain. However, it also introduces a new problem: variance between the contributions made. Variance makes it difficult to produce a coherent document. In this paper, we report on our observations from developing an initial prototype of the open collaboration platform, and derive propositions about how leader-driven open collaboration helps reduce bias while containing variance.
Read the full abstract ▾
Machine Learning
RAILS: Retrieval-Augmented Incremental LLM Clustering at Scale
Armin Oliya, Aleksandra Sawczuk and Radosław Białobrzeski
Using a Large
Language Model (LLM) as the clusterer at production scale is hard:
prompts cannot hold the entire label space, and per-document serial processing does not deliver the throughput real workloads require. We present RAILS, a
retrieval-augmented incremental LLM clusterer that turns clustering into a simple loop over a growing label pool and scales through document batching with bounded concurrency. On six public
benchmarks RAILS exceeds the strongest prior LLM-clustering method on average, lifting accuracy from 51.2% to 59.3%, NMI from 67.2% to 74.8%, and ARI from 45.4% to 54.7%. We further report production-deployment evidence from a SaaS ticket-topic-discovery pipeline, where RAILS has replaced a traditional HDBSCAN stage with higher clustering quality, transparent
prompt-driven control, and stateful incremental operation.
Read the full abstract ▾
Computational Geometry
Witness Set in Weak Visibility Polygons is Polynomial-Time Solvable
Udvas Das, Shouvik Mondal and Sasanka Roy
In the classical Art Gallery Problem (AGP), guards are placed in a polygon so that together they see every point. The Witness Set Problem (WSP), introduced by Amit, Mitchell, and Packer, is a natural dual to the AGP. In this paper, we study the WSP in weak visibility polygons (WVPs), the simple polygons in which every point is seen from some point of one fixed edge. A witness set is a set of points whose visibility regions are pairwise disjoint, so that no single guard sees two of them. A maximum witness set, therefore, lower-bounds the guard number. Exact
polynomial-time
algorithms for the WSP are known only for monotone mountains, a proper subclass of WVPs. We give the first exact
polynomial-time
algorithms for the WSP in WVPs, in two settings. In the Discrete Witness Set Problem (DiscWSP), the witnesses come from a given set of $m$ points, and we find a maximum witness subset in $O(n + m \log(n+m))$ time on an $n$-vertex polygon. The
algorithm rests on a structural fact: the visibility intersection graph of a WVP, in which two points are adjacent if their visibility regions intersect, is a trapezoid graph, that is, an intersection graph of trapezoids between two parallel lines. Moreover, the class of these graphs properly contains the interval graphs and the permutation graphs, which may be of independent interest in graph theory. We also prove an $\Omega(n \log n)$ lower bound in the algebraic decision-tree model for instances with $m = \Theta(n)$, so our
algorithm for DiscWSP is optimum. In the Continuous Witness Set Problem (ContWSP), a witness may be any point of the polygon, and we give an exact
algorithm running in $O(n \log n + \rho^{2}(n + \rho^{2}))$ time, where $\rho$ is the number of reflex vertices.
Read the full abstract ▾
Graphics
Constrained Program Generation for 3D Reaction Animation with a 0.8B Model
Hongyuan Wang, Daming Luo, Nico Pietroni et al. (+2 more)
Visualizing a chemical reaction requires making its molecular changes visible while keeping the animation faithful to the stated chemistry. Equations, structural diagrams and molecular viewers provide complementary descriptions, but assembling an interactive three-dimensional explanation still requires specifying the changes and checking their consistency. We present ChemXRG, a domain-specific
language (DSL) framework that addresses this gap by representing a reaction animation as an executable program. Persistent atom identifiers and explicit bond, charge and grouping operations connect the symbolic reaction to the displayed transformation. This shared representation lets generation, verification and rendering operate on the same account of what changes. Given known, atom-mapped reactant and product structures, reaction-grounded constraints fix input-determined facts and restrict action choices; execution checks validate the resulting transformation before geometry and frames are constructed. We implement this paradigm with a reaction-program corpus and ChemQwen, a trained 0.8B DSL generator. Paired and component evaluations show improved compiler acceptance and normalized full-program agreement under input-conditioned constraints, while identifying remaining failures that require execution checks. A public browser application demonstrates the connection from symbolic reaction descriptions to inspectable programs and interactive 3D animations.
Read the full abstract ▾
All authors Hongyuan Wang, Daming Luo, Nico Pietroni, Christy Liang, Andrew McDonagh
Distributed Computing
Conduit: An Experience Data Plane for Distributed Reinforcement Learning
Sitong Zhang, Tuo Shi, Mario Di Francesco et al. (+2 more)
Distributed reinforcement learning (RL) scales training by parallelizing actors and learners around an Experience Buffer. As RL workloads grow, however, the buffer becomes more than a replay queue: it is the storage substrate of a large-capacity, latency-critical experience path that every iteration traverses to move, transform, sample, and batch experiences before learner updates can begin. Existing RL systems embed this path inside framework control flow or expose it as a request-driven buffer service, leaving experience placement fixed and experience-path work difficult to schedule independently as a runtime-level optimization target. We present Conduit, a framework-agnostic runtime that exposes RL experience management as an explicit systems optimization problem. At its core is the Experience Data Plane (EDP), a runtime abstraction that separates RL experience-handling semantics from framework-specific execution logic by exposing experience ingestion, experience placement, and experience delivery as explicit control points. Built on EDP, Conduit introduces capacity-constrained, bandwidth-aware placement, which distributes experience state across CPU/GPU
memory tiers and nodes under heterogeneous interconnect and device-
memory constraints, and latency-aware scheduling, which controls when experience-path handling runs to reduce exposed experience-path latency while preserving RL semantics. Integrated with RLlib without changing its framework execution logic, Conduit reduces exposed experience-path latency by up to 97% and end-to-end iteration latency by up to 38%, scales to 1,024 GPUs, and preserves convergence.
Read the full abstract ▾
All authors Sitong Zhang, Tuo Shi, Mario Di Francesco, Zeke Wang, Bo Zhao
Computer Vision
MECAIL: Communication-Aware Incremental Learning for Object Detection with 14.6 KB Spatiotemporal Experts
Matthias Neuwirth-Trapp, Maarten Bieshaar, Danda Paudel et al. (+3 more)
Intelligent transportation systems require Incremental Learning (IL) to continually improve their overall performance in dynamic environments. However, most edge devices lack the computational resources to support on-device IL, requiring updates to be transmitted from centralized servers. We propose using this setup to obtain dense, specialized module coverage that adapts a fixed base model to specific spatiotemporal contexts, such as parking lots, gas stations, ferries, or construction sites. However, in order to reliably transmit these modules to the edge device, using TCP, UDP, and BTP over V2X, Wi-Fi, and 2G-5G hardware, we establish a strict limit of 14.6 KB per module to fit within the first TCP window and to minimize UDP/BTP fragmentation. We further introduce Mixture-of-Experts for Communication-Aware Incremental Learning (MECAIL), the first method that meets this strict requirement, in which each new domain or environment is served by a small expert network that adapts the base model. We validate MECAIL on D-RICO and ODinW-13, where it largely matches the performance of parameter-heavy approaches while enabling practical, bandwidth-efficient large-scale deployment. This allows comprehensive coverage by experts for highly specific, focused, and temporary situations.
Read the full abstract ▾
All authors Matthias Neuwirth-Trapp, Maarten Bieshaar, Danda Paudel, Konrad Schindler, Luc Van Gool, Christos Sakaridis
Computational Geometry · Cross-listed from math.OC
The logarithmic spiral is optimal for shoreline search: a computer-assisted proof
Alexander Temerev
A ship starts at a point of the plane and moves at unit speed; it has to reach an unknown straight line, of which neither the distance nor the direction is known. The competitive ratio of a path is the supremum, over all lines, of the time at which the line is reached divided by its distance. Baeza-Yates, Culberson and Rawlins conjectured that a logarithmic spiral, with ratio $C_{\mathrm{sp}} = 13.8111351794611\ldots$, is optimal. We give a computer-assisted proof. Paths are arbitrary: the distance from the start and the polar angle may both decrease. The proof lifts the set of found directions to the universal cover of the circle, where unfolding the polar angle can only increase it (Kneser-Poulsen on the line); a bookkeeping inequality with a monotone final source then bounds the covered measure by the reward of a three-state relaxed control problem, in which inward motion is an ordinary control and excursions below the guaranteed disk are impulses. An explicit $C^1$ storage function, a tensor cubic B-spline plus a closed-form term, satisfies the dissipation inequalities of that problem at the spiral's level and is tight only at the spiral; this is verified with about $10^6$ boxes of Arb ball arithmetic, an exact jet and an interval Hessian at the spiral.
Read the full abstract ▾
Artificial Intelligence
Predicting Postprandial Glycemic Response from Meal Images, Clinical Variables, and Gut Microbiome Information
Varvara Kondratyeva, Kamilia Zaripova, Nassir Navab et al. (+1 more)
Predicting postprandial glycemic response (PPGR) is fundamental to personalized nutrition and type 2 diabetes management, yet existing approaches typically rely on manually reported dietary intake, limiting their scalability in free-living settings. We propose a
multimodal framework that replaces manual dietary logging with image-derived macronutrient estimates and integrates them with clinical variables and gut microbiome information for personalized PPGR prediction. The framework jointly performs image-based macronutrient estimation and glucose prediction, while an
attention-based prediction module models interactions between dietary and host-specific information. We evaluate the proposed approach on a real-world
dataset comprising meal images, continuous glucose monitoring, clinical variables, and gut microbiome profiles. The proposed model outperforms existing PPGR baselines using image-derived nutritional inputs and approaches the performance of methods that rely on manually reported macronutrients despite using automatically estimated nutritional information. These results demonstrate that combining image-derived nutrition with complementary clinical and gut microbiome information provides a practical foundation for scalable personalized PPGR prediction.
Read the full abstract ▾
All authors Varvara Kondratyeva, Kamilia Zaripova, Nassir Navab, Azade Farshad
Computer Vision
Do LiDAR Language Models Really Understand Spatio-temporal Relationships?
Runyi Yang, Murat Akkoyun, Di Wen et al. (+8 more)
Recent 4D
LiDAR language models aim to reason about objects and their evolving spatial relationships. Yet, in our evaluation, always selecting the same option nearly matches the multiple-choice accuracy of two B4DL-derived configurations. We introduce
LiDAR-Hallu, a geometry-referenced
benchmark and diagnostic protocol with 10,000 questions across 150 nuScenes scenes. It covers object existence, ego-relative position, distance ordering, relative motion, and temporal localization, with explicit rules for selecting objects, comparing times, and determining reference answers. Our protocol combines fixed-answer and candidate-content controls, cross-scene pairs with identical
prompts but opposite reference answers, and relation-specific recall. Analysis of 100,000 recorded responses reveals failures hidden by aggregate accuracy. Candidate duration alone makes temporal answers predictable without observing
LiDAR. On paired questions, the models frequently give the same answer to scenes requiring opposite answers. Relation-specific analysis further shows that both configurations miss every positive lateral-motion case across all tested conditions. Temporal-shuffle contrastive decoding provides little net improvement, as repairs are largely offset by new errors and the main failures persist. These results show that evaluating spatio-temporal
reasoning requires testing whether models distinguish the queried physical relationships, rather than relying on individual-answer accuracy alone. The source code, checkpoints, and data are released at https://github.com/Awesome4D/4DMLLM_
Hallucination_Bench.
Read the full abstract ▾
All authors Runyi Yang, Murat Akkoyun, Di Wen, Ruiping Liu, Yufan Chen, Junwei Zheng, Xiaoye Wang, Kailun Yang, Danda Pani Paudel, Luc Van Gool, Kunyu Peng
Numerical Analysis · Cross-listed from math.OC
Numerical approximation of distributed optimal control problems governed by linear thermo-poroelasticity
Jeonghun J. Lee
In this paper, we propose and analyze a novel three-field symmetric formulation for optimal control problems governed by linear thermo-poroelasticity. The state variables are the solid displacement $\boldsymbol{u}$, the fluid pressure $p$, and the temperature $\theta$, and both the distributed fluid source $m_p$ and the distributed heat source $m_\theta$ serve as control variables. We prove well-posedness of the symmetric differentiated formulation, establish the existence and uniqueness of an optimal control pair $(\bar{m}_p,\bar{m}_\theta)$, and derive the first-order necessary optimality conditions via a coupled three-field adjoint system. For the fully discrete scheme based on dG(0) time discretization, we employ a variational discretization approach for both controls and derive a priori error estimates of order $\mathcal{O}(h^s + \Delta t)$ for the state, adjoint, and control errors. We also present manufactured-solution experiments that validate the fully discrete optimality system and illustrate the storage-degenerate regime covered by our analysis.
Read the full abstract ▾
Cryptography and Security
ActGov: Governing LLM Agent Actions via Policy-Constrained Validation
Kaiyuan Zhang, Yuke Peng, Ke Jiang et al. (+1 more)
Large
language model (LLM)
agents increasingly execute long-horizon workflows through external tools, allowing untrusted outputs to influence subsequent actions and exceed user authorization. Existing defenses isolate injected content or constrain execution with predefined plans and static policies, but these approaches are brittle under dynamic workflows and scale poorly across extensible tool ecosystems. In this work, we present ActGov, a runtime enforcement framework that validates each LLM-proposed tool action before it causes external effects. Built on a unified semantic model of authorization, actions, runtime context, and security constraints, the ActGov-Policy component iteratively constructs a policy set from tool specifications, benign tasks, and observed failure traces, with each update verified through SMT-based counterexample checking. At runtime, ActGov-Runtime abstracts each tool call into finite policy records and permits it only if it remains within the task-scoped authorization boundary and satisfies all applicable policies. This per-action enforcement preserves authorization throughout long-horizon, dynamically branching workflows. We evaluate ActGov on the AgentDojo and AgentDyn
benchmarks across multiple models and attack configurations. It shows that ActGov consistently reduces the success rate of indirect
prompt-injection attacks while preserving task utility, significantly outperforming existing defenses. These results demonstrate that ActGov can enforce fine-grained authorization over dynamic
agent executions without relying on the underlying LLM to correctly identify malicious instructions.
Read the full abstract ▾
All authors Kaiyuan Zhang, Yuke Peng, Ke Jiang, Yinqian Zhang
Machine Learning
WPBench: A Comprehensive Benchmark for Wind Power Forecasting
Yuhan Zhu, Jilin Hu, Xinying Cai et al. (+7 more)
Accurate, reliable, and deployable wind power
forecasting is critical for power system dispatch, renewable energy integration, and electricity
market operations. Progress in this field hinges on the ability to empirically and comprehensively
benchmark forecasting methods. Yet existing
benchmarks fall short of supporting systematic evaluation in four key aspects: 1) limited coverage of wind power scenarios across turbine scale, variable composition, and spatial structure; 2) incomplete coverage of
forecasting model families; 3) evaluation metrics misaligned with wind power requirements; and 4) limited structure-aware diagnostics beyond individual temporal patterns. To address these limitations, we propose WPBench, a comprehensive, fair, and extensible
benchmark for wind power
forecasting. WPBench integrates 26 public
datasets organized by turbine scale and variable composition, spanning single-turbine, multi-turbine, univariate, and multivariate settings. Under unified processing, training, and evaluation protocols, it
benchmarks 19 representative models covering traditional methods, deep temporal models, spatio-temporal models, and foundation models. Beyond point-wise errors, WPBench assesses forecast-curve fidelity and computational efficiency, and delivers structure-aware diagnostics across temporal, variable-dependency, and spatial-dependency perspectives. Together, these capabilities enable systematic model comparison across diverse wind scenarios and provide a reusable platform for future research.
Read the full abstract ▾
All authors Yuhan Zhu, Jilin Hu, Xinying Cai, Yingshan Li, Li Ma, Xiangfei Qiu Linsen Li, Kai Zhang, Yao Fu, Weihao Jiang, Bin Yang
Machine Learning · Cross-listed from physics.plasm-ph
Horizon-Aware Early Event Prediction for Tokamak Disruption Alarms
Takeshi Koshizuka and Takaharu Yaguchi
Disruption predictors must warn the
plasma control system early enough for the intended response, and the required warning time differs between mitigation and avoidance. We view a warning system as a predictor that outputs a disruption risk and an alarm policy that acts on it, and ask which risk the predictor should learn. Binary classifiers do not specify when a disruption occurs, and survival analysis models the full distribution of the time to disruption, whereas an alarm depends only on the risk within a finite horizon. We introduce two horizon-focused targets into tokamak disruption prediction: a horizon-truncated survival objective and fixed-horizon early event prediction (EEP) with temporal label smoothing (TLS). We compare them with binary classification and survival analysis on the DIII-D, Alcator C-Mod, and EAST databases of DisruptionBench, evaluating the risk by its threshold-free ranking at a 40 ms deadline and the alarm policy by thresholds calibrated on validation data. On every device, TLS outperforms the deep survival models in both ranking and alarm performance; TLS performs best on DIII-D and EAST, and a random forest trained on a hard horizon target performs best on C-Mod. TLS also improves the calibrated alarm performance over hard targets placed at either end of its smoothing range on all devices, isolating the benefit of label smoothing. Horizon truncation alone did not consistently improve survival analysis. Alarm performance remains limited on C-Mod and DIII-D and varies strongly across training seeds, and ablations show that the target should extend to the required warning time.
Read the full abstract ▾
Machine Learning
MUSE: Dependency-Aware Adaptation of a Frozen Vision Backbone for Multivariate Time Series Forecasting
Xinying Cai, Junkai Lu, Yuhan Zhu et al. (+3 more)
Multivariate time-series
forecasting is essential to many real-world applications. Recent large
vision models (LVMs) offer a promising paradigm by transferring cross-domain visual priors to time-series
forecasting. However, existing LVM-based methods face two key challenges: balancing independent visual representation spaces with cross-variable dependency modeling, and adapting
vision backbones pretrained on natural images to the distinct temporal semantics of time-series images. To address these challenges, we propose MUSE, a dependency-aware adaptation framework built on a fully frozen pretrained MAE. First, the Variable Context Refinement Module (VCR) aggregates shared temporal information within each variable and models cross-variable contextual dependencies while preserving independent visual spaces. Second, the Temporal-Periodic Refinement Module (TPR) performs lightweight refinement at different encoder depths and explicitly models across-period temporal dependencies and within-period periodic dependencies. The two modules independently produce forecasts, which are fused through a learnable prediction-level gate. Experiments on 10 real-world
datasets demonstrate that MUSE achieves state-of-the-art performance.
Read the full abstract ▾
All authors Xinying Cai, Junkai Lu, Yuhan Zhu, Xiaoyun Yu, Xiangfei Qiu, Jilin Hu
Machine Learning
Comparing Latent Concept Formation in State Space Models and Transformers via Sparse Autoencoders
Rithin Nagaraj, Rupa Laalasa Oruganti, Prerna Subhashchandra Kunder et al. (+1 more)
The quadratic scaling of
Transformer self-
attention has driven the adoption of sub-quadratic Selective State Space Models (SSMs) like Mamba, which compress past context into a fixed-size recurrent hidden state. This strict informational bottleneck raises a foundational question for mechanistic interpretability: do SSMs and
Transformers learn fundamentally distinct latent representations? In this work, we employ Sparse Autoencoders (SAEs) to conduct a large-scale, feature-level correspondence analysis between Mamba-130m and Pythia-70m over a 10-million
token corpus. Contrary to hypotheses predicting widespread architectural divergence, we find no evidence of systematic representational divergence between architectures: across the observed Jaccard distribution, 99.98% of Mamba features cluster toward the upper
alignment boundary, providing preliminary feature-level support for the Universality Hypothesis. We further identify and qualitatively characterize this microscopic fraction (0.02%) of diverging features, finding patterns consistent with the hypothesis that the recurrent bottleneck selectively limits the parsing of rigid syntax rather than broad semantic ontology. We demonstrate that while Pythia's unconstrained
attention permits the monosemantic decomposition of distinct formatting edge-cases, Mamba is forced to compress unrelated syntactical anomalies into polysemantic "junk drawer"
neurons to preserve state capacity. Collectively, these results suggest that architectural routing mechanisms may have negligible impact on core semantic understanding, with representational divergence confined to extreme structural margins.
Read the full abstract ▾
All authors Rithin Nagaraj, Rupa Laalasa Oruganti, Prerna Subhashchandra Kunder, Ashwini M Joshi
Programming Languages
Categorical Message Passing Language (CaMPL): Syntax and Semantics
Robin Cockett, Daniel Kiyoshi Hashimoto, Alexanna Little Berg et al. (+1 more)
We introduce a novel functional-style concurrent programming
language called Categorical Message Passing
Language (CaMPL) which is designed using the mathematics of linear actegories. This mathematical underpinning gives CaMPL programs useful properties such as deadlock freedom, and additionally, livelock freedom for programs without general recursive processes. We explore CaMPL's type system through a series of code examples. The current proto-alpha version of the compiler and the abstract machine is implemented in Haskell. A reader is encouraged to experiment with writing CaMPL programs either using the online compiler https://campl-app.vercel.app/ or by installing CaMPL from https://campl-ucalgary.github.io – our website has detailed instructions on how to run CaMPL code.
Read the full abstract ▾
All authors Robin Cockett, Daniel Kiyoshi Hashimoto, Alexanna Little Berg, Priyaa Varshinee Srinivasan
Computational Engineering
Adaptive Parallel-in-Time Integration with Dynamic Resource Management
Keerthi Gaddameedi, Dominik Huber, Martin Schreiber et al. (+4 more)
As computational resources continue to grow, the strong-scaling limitations of spatial parallelism motivate the pursuit of additional concurrency in the temporal dimension, particularly for applications with hard time constraints, such as weather and
climate simulations. The Parallel Full Approximation Scheme in Space and Time (PFASST) is a parallel-in-time method based on Spectral Deferred Corrections (SDC). It computes multiple timesteps concurrently by coupling fine- and coarse-grid SDC sweeps using multigrid Full Approximation Scheme (FAS) corrections. However, PFASST's convergence is often problem-dependent, demanding a variable number of parallel timesteps and, hence, computing resources at different times throughout the simulation. Dynamic Resource Management (DRM) provides a remedy for this challenge by enabling the adaptive adjustment of computational resources and algorithmic parameters at runtime. In this work, we present our novel approach to extending PFASST with DRM, which enables (a) dynamic adaptation of computing resources, (b) adaptive selection of the number of PFASST iterations based on local convergence behavior, and (c) coupling of these two adaptations into a single resizing strategy. With this approach, we demonstrate for the first time that optimal configurations can be identified in real time for each application, rather than relying on static allocation. Furthermore, we show that convergence-informed tuning of PFASST improves resource utilization and convergence efficiency.
Read the full abstract ▾
All authors Keerthi Gaddameedi, Dominik Huber, Martin Schreiber, Hans-Joachim Bungartz, Valentina Schüller, Tobias Neckel, Martin Schulz
Robotics
FoldQuantVLA: Native Low-Bit Quantization of Vision-Language-Action Models via Consistent Folding
Hung T. Ho, Khanh D. Nguyen, Quang D. Nguyen et al. (+5 more)
Low-bit
vision-
language-action inference must reduce observation-to-action latency while preserving
robot behavior. We present FoldQuantVLA, a post-training quantization framework that carries a consistent activation representation through calibration, weight rounding, and native integer execution. It combines channel scaling and block Hadamard transforms with dynamic per-
token quantization, without policy retraining. Custom TensorRT plugins execute projections in both the
language backbone and iterative action expert with four-bit weights and activations (W4A4) on Ada GPUs and Jetson AGX Orin. Evaluation spans LIBERO, SimplerEnv, and two
robot platforms. Across three GR00T checkpoints and $\pi_{0.5}$, W4A4 achieves $1.20$ to $1.33\times$ speedups over floating-point TensorRT on Orin and $1.25$ to $1.52\times$ on desktop. Retaining
language attention-output and feed-forward down projections at eight bits (W8A8) improves held-out action fidelity on all four checkpoints. Across four real-
robot tasks, this configuration raises observed GR00T N1.7 success from $80.0\%$ with uniform W4A4 to $92.5\%$ over 80 trials per configuration, with a measured additional Orin latency of 1 ms.
Read the full abstract ▾
All authors Hung T. Ho, Khanh D. Nguyen, Quang D. Nguyen, Thanh Q. Duong, Ngan Le, Meng Guo, Vien A. Ngo, An T. Le
Machine Learning
1% of Tokens Can Be Enough: On Gradient Estimation in On-Policy Distillation
Huanxin Sheng, Zhiling Ye, Haonan Wang et al. (+3 more)
Sparse on-policy
distillation (OPD) allocates teacher supervision to a small subset of
tokens in student-generated trajectories. However, useful teacher guidance can yield a noisy update when its gradient is estimated from a sampled next
token. We study this estimation problem at a fixed prefix in information geometry and propose an information-efficiency ratio (IER) based on a signal-to-noise decomposition. IER characterizes relative gradient estimation error under an optimal scalar baseline. A candidate-set approximation enables
token selection based on IER and its combination with existing usefulness scores, while retaining the sampled reverse-KL training objective. On mathematical and medical
reasoning tasks, adding IER improves existing selectors in multiple settings, with sparse configurations matching or exceeding full OPD without
token selection at small
token budgets of 0.1%–1%. These results support accounting for both usefulness and gradient-estimation reliability when allocating sparse supervision. Our code is available at https://github.com/BruceSheng1202/IER-OPD.
Read the full abstract ▾
All authors Huanxin Sheng, Zhiling Ye, Haonan Wang, Jian Wang, Jinjie Gu, Jian Kang
Information Retrieval
What Makes a Good Semantic ID for Generative Recommendation? A Reproducibility Study
Yufei Chen, Junchen Fu, Jujia Zhao et al. (+2 more)
Generative recommendation has emerged as an active research direction, where items are commonly represented by semantic IDs (SIDs): discrete codes generated
token by
token. Despite strong empirical results, SID designs vary widely in construction strategy, codebook organization, and code length, making their true impact on recommendation performance unclear. We conduct a large-scale reproducibility study to systematically investigate the impact of semantic ID design on generative recommendation under a unified experimental framework. We focus on a fundamental question: What makes a good semantic ID for generative recommendation? To answer this question, we examine four aspects: the relative effectiveness of different semantic ID designs, the connection between codebook utilization and recommendation quality, the effect of semantic code length, and the influence of semantic ID design on local item semantic preservation. Through a unified evaluation and additional cross-
dataset controlled analyses, we find that the effects of SID design are largely non-monotonic: no single SID design is universally best, and commonly used RQ-VAE- and OPQ-based designs can behave inconsistently across
datasets. The method with the most balanced first-level codebook is not consistently the best recommender, showing that utilization is diagnostic but insufficient. Scaling either the generative backbone or the SID length is also not always beneficial. Finally, semantic-neighborhood analysis reveals that no single SID design dominates all notions of local semantic preservation; instead, different designs exhibit complementary strengths that remain stable across
datasets and neighborhood sizes. Our study provides a controlled and reproducible understanding of semantic ID design and offers practical insights for future generative recommender systems.
Read the full abstract ▾
All authors Yufei Chen, Junchen Fu, Jujia Zhao, Yukun Zhao, Zhaochun Ren
Computer Vision
Estimating Accurate Hand Pose in Camera Space with Vision Transformer
Kaiwen Ren, Yiran Jiang, Yongjing Ye et al. (+1 more)
Monocular RGB-based hand pose estimation has emerged as a critical research frontier in computer
vision. The local hand pose estimation methods predict hand poses relative to the wrist, while global hand pose estimation also requires estimating the wrist's position in the camera coordinate system. However, this camera-space estimation confronts two fundamental challenges: (1) depth ambiguity in
monocular settings, and (2) the coupling effect of hand local poses and global wrist positions in the perspective projections. In particular, this coupling reflects that the projections are jointly determined by local hand poses, wrist positions, and camera intrinsics. To overcome these challenges, our framework proposes two key innovations: Transformation-Isomorphism Supervision for hand-depth information extraction and Perspective Information
Embedding for resolving above coupling effect of local pose and wrist position, both integrated within the mainstream encoder-decoder architecture. Besides, we propose a novel framerate-aware multi-
dataset training strategy for sequential pose refinement. Our fully integrated approach achieves at most 37.1% superiority in CS-MJE over SOTA on HO3D. Project page: https://github.com/Mine268/CS-ViT.
Read the full abstract ▾
All authors Kaiwen Ren, Yiran Jiang, Yongjing Ye, Shihong Xia
Machine Learning · Cross-listed from math.OC
Complexities of Weak Proximal Oracle Methods for Composite Convex Optimization
Dan Garber
We consider a standard convex composite optimization problem with either smooth or nonsmooth objective function, and under quadratic growth. In recent years, several works gave
algorithms based on a weak proximal oracle (WPO) that essentially match in oracle complexities proximal (sub)gradient methods relying on exact prox operations. Importantly, such WPOs, which relax the strong optimality condition of the standard prox operator, may admit much more efficient implementation in terms of runtime when optimal solutions have some sparse structure. A question remained if such WPO-based methods can be accelerated (in the sense of Nesterov's accelerated gradient). In this work we provide a negative answer by establishing lower bounds against both deterministic and randomized methods. Thus, while WPOs can substantially reduce the cost of individual oracle calls, this comes with an inherent loss in oracle complexity. We also provide a new upper-bound for WPO-based nonsmooth convex composite optimization, nearly matching the proximal subgradient method.
Read the full abstract ▾