Vision-Language-Action (VLA) models offer a promising autonomous driving paradigm for leveraging world knowledge and reasoning capabilities, especially in long-tail scenarios. However, existing VLA models often struggle with the high latency in action generation using an autoregressive generation framework and exhibit limited robustness. In this paper, we propose SpanVLA, a novel end-to-end autonomous driving framework, integrating an autoregressive reasoning and a flow-matching action expert. First, SpanVLA introduces an efficient bridge to leverage the vision and reasoning guidance of VLM to efficiently plan future trajectories using a flow-matching policy conditioned on historical trajectory initialization, which significantly reduces inference time. Second, to further improve the performance and robustness of the SpanVLA model, we propose a GRPO-based post-training method to enable the VLA model not only to learn from positive driving samples but also to learn how to avoid the typical negative behaviors and learn recovery behaviors. We further introduce mReasoning, a new real-world driving reasoning dataset, focusing on complex, reasoning-demanding scenarios and negative-recovery samples. Extensive experiments on the NAVSIM (v1 and v2) demonstrate the competitive performance of the SpanVLA model. Additionally, the qualitative results across diverse scenarios highlight the planning performance and robustness of our model.
We propose SmoothCruiser, a new planning algorithm for estimating the value function in entropy-regularized Markov decision processes and two-player games, given a generative model of the environment. SmoothCruiser makes use of the smoothness of the Bellman operator promoted by the regularization to achieve problem-independent sample complexity of order O~(1/epsilon^4) for a desired accuracy epsilon, whereas for non-regularized settings there are no known algorithms with guaranteed polynomial sample complexity in the worst case.
Understanding artworks requires multi-step reasoning over visual content and cultural, historical, and stylistic context. While recent multimodal large language models show promise in artwork explanation, they rely on implicit reasoning and internalized knowl- edge, limiting interpretability and explicit evidence grounding. We propose A-MAR, an Agent-based Multimodal Art Retrieval framework that explicitly conditions retrieval on structured reasoning plans. Given an artwork and a user query, A-MAR first decomposes the task into a structured reasoning plan that specifies the goals and evidence requirements for each step. Retrieval is then conditionedon this plan, enabling targeted evidence selection and supporting step-wise, grounded explanations. To evaluate agent-based multi- modal reasoning within the art domain, we introduce ArtCoT-QA. This diagnostic benchmark features multi-step reasoning chains for diverse art-related queries, enabling a granular analysis that extends beyond simple final answer accuracy. Experiments on SemArt and Artpedia show that A-MAR consistently outperforms static, non planned retrieval and strong MLLM baselines in final explanation quality, while evaluations on ArtCoT-QA further demonstrate its advantages in evidence grounding and multi-step reasoning ability. These results highlight the importance of reasoning-conditioned retrieval for knowledge-intensive multimodal understanding and position A-MAR as a step toward interpretable, goal-driven AI systems, with particular relevance to cultural industries. The code and data are available at: https://github.com/ShuaiWang97/A-MAR.
Effective human-robot teaming is crucial for the practical deployment of robots in human workspaces. However, optimizing joint human-robot plans remains a challenge due to the difficulty of modeling individualized human capabilities and preferences. While prior research has leveraged the multi-cycle structure of domains like manufacturing to learn an individual's tendencies and adapt plans over repeated interactions, these techniques typically consider task-level and motion-level adaptation in isolation. Task-level methods optimize allocation and scheduling but often ignore spatial interference in close-proximity scenarios; conversely, motion-level methods focus on collision avoidance while ignoring the broader task context. This paper introduces RAPIDDS, a framework that unifies these approaches by modeling an individual's spatial behavior (motion paths) and temporal behavior (time required to complete tasks) over multiple cycles. RAPIDDS then jointly adapts task schedules and steers diffusion models of robot motions to maximize efficiency and minimize proximity accounting for these individualized models. We demonstrate the importance of this dual adaptation through an ablation study in simulation and a physical robot scenario using a 7-DOF robot arm. Finally, we present a user study (n=32) showing significant plan improvement compared to non-adaptive systems across both objective metrics, such as efficiency and proximity, and subjective measures, including fluency and user preference. See this paper's companion video at: https://youtu.be/55Q3lq1fINs.
Human mobility data are used in numerous applications, ranging from public health to urban planning. Human mobility is inherently sensitive, as it can contain information such as religious beliefs and political affiliations. Historically, it has been proposed to modify the information using techniques such as aggregation, obfuscation, or noise addition, to adequately protect privacy and eliminate concerns. As these methods come at a great cost in utility, new methods leveraging development in generative models, were introduced. The extent to which such methods answer the privacy-utility trade-off remains an open problem. In this paper, we introduced a first step towards solving it, by the introduction and application of a new framework for utility evaluation. Furthermore, we provide evidence that privacy evaluation remains a great challenge to consider and that it should be tackled through adversarial evaluation in accordance with the current EU regulation. We propose a new membership inference attack against a subcategory of generative models, even though this subcategory was deemed private due to its resistance over the trajectory user-linking problem.
Air Traffic Flow Management (ATFM) traffic regulations are being increasingly used as rising demand meets persistent workforce shortages. This operational strain has amplified a critical phenomenon that we call \emph{regulation cascading}: the compounding, non-linear interactions that occur when multiple regulations influence one another in unpredictable ways. As the number and complexity of regulations grow, cascading effects become more pronounced, undermining the network operator's ability to protect sectors reliably. To address this challenge, we introduce Regulation Zero 2, an updated sequential planning framework that natively operates in the regulation space, optimizing over ordered sequences of flow-level regulations that remain compatible as much as possible with existing slot-allocation systems such as CASA and RBS++. We equipped Regulation Zero 2 with new heuristics to render flow finding more efficient. At its core, the method employs a hierarchical Monte Carlo Tree Search (MCTS) that first samples congestion hotspots and then selects candidate regulations synthesized by a local proposal engine. Each proposal is evaluated by a fast First-Planned-First-Served (FPFS) allocator to estimate its reward, with these feedbacks guiding the subsequent MCTS exploration. Experiments on many pan-European summer-peak traffic days that Regulation Zero delivers promising and consistent performance. Compared to a flight-centric simulated-annealing and NSGA-II baselines, it achieves markedly higher objective improvements, while maintaining a tighter scope of impact on the network. Ablation studies also found that Regulation Cascading could reduce up to 50\% of potential effectiveness. RZ also preserves FPFS fairness and supports expert knowledge injection, offering a pragmatic and low-disruption pathway toward automation in operations.
Multimodal Large Language Models are increasingly adopted as autonomous agents in interactive environments, yet their ability to proactively address safety hazards remains insufficient. We introduce SafetyALFRED, built upon the embodied agent benchmark ALFRED, augmented with six categories of real-world kitchen hazards. While existing safety evaluations focus on hazard recognition through disembodied question answering (QA) settings, we evaluate eleven state-of-the-art models from the Qwen, Gemma, and Gemini families on not only hazard recognition, but also active risk mitigation through embodied planning. Our experimental results reveal a significant alignment gap: while models can accurately recognize hazards in QA settings, average mitigation success rates for these hazards are low in comparison. Our findings demonstrate that static evaluations through QA are insufficient for physical safety, thus we advocate for a paradigm shift toward benchmarks that prioritize corrective actions in embodied contexts. We open-source our code and dataset under https://github.com/sled-group/SafetyALFRED.git
Wireless agentic systems enable agents to autonomously perceive, reason, and act. However, existing works neglect the tight coupling between sensing and control in closed-loop integrated sensing and communication (ISAC) systems. In this paper, we propose an active inference (AIF)-driven wireless agentic system for closed-loop ISAC, which jointly optimizes control and sensing resource allocation via backward--forward message passing on a factor graph. The AIF agent maintains a generative model as a digital twin by integrating a localization model for uncertainty-aware state inference and a localization channel knowledge map (CKM) for approximating observation quality during planning. Simulation results demonstrate that the AIF-enabled agent adaptively allocates sensing resources based on spatially varying channel conditions, achieving superior balance among tracking accuracy, control effort, and sensing resource consumption over baseline strategies.
Current 3D mapping pipelines generally assume static environments, which limits their ability to accurately capture and reconstruct moving objects. To address this limitation, we introduce the novel task of active mapping of moving objects, in which a mapping agent must plan its trajectory while compensating for the object's motion. Our approach, Paparazzo, provides a learning-free solution that robustly predicts the target's trajectory and identifies the most informative viewpoints from which to observe it, to plan its own path. We also contribute a comprehensive benchmark designed for this new task. Through extensive experiments, we show that Paparazzo significantly improves 3D reconstruction completeness and accuracy compared to several strong baselines, marking an important step toward dynamic scene understanding. Project page: https://davidea97.github.io/paparazzo-page/
Desktop interaction streams provide a continuous, privacy-sensitive record of interleaved user tasks. Transforming these streams into task-organized personal logs on-device faces two main challenges: exhaustive Vision-Language Model (VLM) processing strains local resources, and global stream processing causes cross-task context pollution. We present FOCAL (Filtered On-device Continuous Activity Logging), a privacy-first multi-agent system utilizing a unified filter-plan-log architecture. It cascades a lightweight Filter Agent for noise suppression, a text-only Brain Agent for task attribution, a Record Agent for selective visual reasoning, and a task-isolated Memory Agent for context-coherent summarization. Experiments on DesktopBench (comprising 2,572 screenshots across 420 complex sessions) show FOCAL reduces total token consumption by 60.4% and VLM call count by 72.3% versus a baseline, while boosting Key Information Recall (KIR) from 0.38 to 0.61. Crucially, under $A{\to}B{\to}A$ task interruptions, FOCAL maintains Task Acc 0.81 and KIR 0.80, whereas the baseline collapses to Task Acc 0.03. FOCAL pioneers the efficient, on-device summarization of instruction-free desktop streams into multi-perspective personal logs.
For operational regional synthetic aperture radar (SAR) reconnaissance, mission success depends not only on geometric visibility but also on whether geometric feasibility, prescribed imaging quality, and timely data delivery can be met together within the planning horizon. This paper develops an effective window framework for regional SAR window generation, per window signal level quality screening, and hybrid direct-relay closed loop scheduling. Through coarse angular bandpass screening, a planar characteristic curve containment test, and one dimensional boundary bisection, the framework forms geometry feasible candidate observation windows with millisecond-level accuracy for their entry and exit times. Each candidate window is then assessed in stripmap mode with a companion point target under a unified echo generation and Back Projection (BP) imaging workflow; only windows whose range and azimuth impulse response width (IRW), peak sidelobe ratio (PSLR), and integrated sidelobe ratio (ISLR) all satisfy the preset thresholds are retained. The retained observation, relay, and downlink windows feed a quality constrained hybrid direct-relay closed-loop mixed-integer linear programming (MILP) formulation for joint scheduling of observation and ground return. Numerical experiments confirm millisecond-level agreement with STK reference timing for window boundaries. Every candidate window is screened against preset imaging quality thresholds. Hybrid closed-loop scheduling improves closure performance and ground returned data volume relative to a relay-only baseline
Generative engines (GEs) are reshaping information access by replacing ranked links with citation-grounded answers, yet current Generative Engine Optimization (GEO) methods optimize each instance in isolation, unable to accumulate or transfer effective strategies across tasks and engines. We reframe GEO as a strategy learning problem and propose MAGEO, a multi-agent framework in which coordinated planning, editing, and fidelity-aware evaluation serve as the execution layer, while validated editing patterns are progressively distilled into reusable, engine-specific optimization skills. To enable controlled assessment, we introduce a Twin Branch Evaluation Protocol for causal attribution of content edits and DSV-CF, a dual-axis metric that unifies semantic visibility with attribution accuracy. We further release MSME-GEO-Bench, a multi-scenario, multi-engine benchmark grounded in real-world queries. Experiments on three mainstream engines show that MAGEO substantially outperforms heuristic baselines in both visibility and citation fidelity, with ablations confirming that engine-specific preference modeling and strategy reuse are central to these gains, suggesting a scalable learning-driven paradigm for trustworthy GEO. Code is available at https://github.com/Wu-beining/MAGEO
We present an update on the ongoing computation of the isospin-breaking effects in the Pion Decay Constant from the BMW Collaboration. The calculation is carried out with N$_f$=2+1+1 staggered quarks with a near-physical pion mass and QED$_{\text{L}}$. We give an update on the isosymmetric value and the current determination used to compute the gradient-flow scale $w_{0}$, then we present some preliminary results of the valence-valence contribution to the axial-pseudoscalar correlator for different volumes and lattice spacings. We also discuss the next steps and plans.
Wireless links deployed in orchards often exhibit significant variability in the strength of the received signal that is not adequately captured by classical distance-based propagation models. In row-structured olive groves, signal attenuation differs markedly between along-row and cross-row propagation directions, leading to discrepancies when using omnidirectional propagation assumptions such as those adopted in the Free Space Path Loss (FSPL) model or ITU-R vegetation loss formulations. This paper proposes a topology-based propagation model that explicitly accounts for orchard layout and the relative positions of radio devices within the plantation structure. Experimental validation was conducted using LoRa technology operating at 868 MHz, and the results were compared with established models from the literature and with the proposed two-dimensional model. The proposed approach achieves a closer fit to measured RSSI data than conventional models, providing a more reliable basis for link budgeting and network planning in structured agricultural environments.
Knowledge work demands sustained self-regulation, prioritization, and reflection-yet existing planning tools only partially support these needs. Digital to-do list applications feature task persistence but lack goal representation. Paper-based planning frameworks offer effective planning strategies but cannot adapt to individual users. Conversational AI systems enable flexible reflection but lack persistence and accountability. Moreover, none of these tools address a fundamental challenge: users' expressed demands often diverge from their underlying needs. This paper introduces seneca, a conceptual framework for a personalized, AI-assisted planner that integrates the complementary strengths of these three approaches. seneca combines a conversational agent that scaffolds reflection and asks clarifying questions, a persistent database that tracks goals and behavioral patterns, and a processor that synchronizes information between them. We describe this architecture and outline a phased evaluation strategy combining automated testing with simulated users and longitudinal human studies measuring goal attainment, planning realism, and goal-value alignment.
Understanding road scenes in a geometrically consistent, scene-centric representation is crucial for planning and mapping. We present GOLD-BEV, a framework that learns dense bird's-eye-view (BEV) semantic environment maps-including dynamic agents-from ego-centric sensors, using time-synchronized aerial imagery as supervision only during training. BEV-aligned aerial crops provide an intuitive target space, enabling dense semantic annotation with minimal manual effort and avoiding the ambiguity of ego-only BEV labeling. Crucially, strict aerial-ground synchronization allows overhead observations to supervise moving traffic participants and mitigates the temporal inconsistencies inherent to non-synchronized overhead sources. To obtain scalable dense targets, we generate BEV pseudo-labels using domain-adapted aerial teachers, and jointly train BEV segmentation with optional pseudo-aerial BEV reconstruction for interpretability. Finally, we extend beyond aerial coverage by learning to synthesize pseudo-aerial BEV images from ego sensors, which support lightweight human annotation and uncertainty-aware pseudo-labeling on unlabeled drives.
Satellite image composition plays a critical role in remote sensing applications such as data augmentation, disaste simulation, and urban planning. We propose HarmoniDiff-RS, a training-free diffusion-based framework for harmonizing composite satellite images under diverse domain conditions. Our method aligns the source and target domains through a Latent Mean Shift operation that transfers radiometric characteristics between them. To balance harmonization and content preservation, we introduce a Timestep-wise Latent Fusion strategy by leveraging early inverted latents for high harmonization and late latents for semantic consistency to generate a set of composite candidates. A lightweight harmony classifier is trained to further automatically select the most coherent result among them. We also construct RSIC-H, a benchmark dataset for satellite image harmonization derived from fMoW, providing 500 paired composition samples. Experiments demonstrate that our method effectively performs satellite image composition, showing strong potential for scalable remote-sensing synthesis and simulation tasks. Code is available at: https://github.com/XiaoqiZhuang/HarmoniDiff-RS.
Recent advances in Text-To-Speech (TTS) synthesis have seen the popularity of multi-stage approaches that first predict semantic tokens and then generate acoustic tokens. In this paper, we extend the coarse-to-fine generation paradigm to the temporal domain and introduce Chain-of-Details (CoD), a novel framework that explicitly models temporal coarse-to-fine dynamics in speech generation using a cascaded architecture. Our method progressively refines temporal details across multiple stages, with each stage targeting a specific temporal granularity. All temporal detail predictions are performed using a shared decoder, enabling efficient parameter utilization across different temporal resolutions. Notably, we observe that the lowest detail level naturally performs phonetic planning without the need for an explicit phoneme duration predictor. We evaluate our method on several datasets and compare it against several baselines. Experimental results show that CoD achieves competitive performance with significantly fewer parameters than existing approaches. Our findings demonstrate that explicit modeling of temporal dynamics with the CoD framework leads to more natural speech synthesis.
The Traveling Thief Problem (TTP) is a multi-component optimization problem that captures the interplay between routing and packing decisions by combining the classical Traveling Salesperson Problem (TSP) and the Knapsack Problem (KP). The TTP has gained significant attention in the evolutionary computation literature and a wide range of approaches have been developed over the last 10 years. Judging the performance of these algorithms in particular in terms of how close the get to optimal solutions is a very challenging task as effective exact methods are not available due to the highly challenging traveling component. In this paper, we study the tour-optimization component of TTP under a fixed packing plan. We formulate this task as a weighted variant of the TSP, where travel costs depend on the cumulative weight of collected items, and investigate how different distance metrics and cost functions affect computational complexity. We present an $(O(n^2))$-time dynamic programming algorithm for the path metric with general cost functions, prove that the problem is NP-hard even on a star metric, and develop constant-factor approximation algorithms for star metrics. Finally, we also develop an approximation algorithm for the problem under a general metric with a linear cost function. We complement our theoretical results with experimental evaluations on standard TTP instances adjusted to a path metric. Our experimental results demonstrate the practical effectiveness of our approaches by comparing it to solutions produced by popular iterative search algorithms. The results show that our methods are able to significantly improve the quality of solutions for some benchmark instances by optimizing the traveling part while pointing out the optimality of the travel component for other solutions obtained by iterative search methods.
Developing Visual Analytics (VA) applications requires integrating complex machine learning models with expressive interactive interfaces. Developers face a stark trade-off: building tightly-coupled monoliths plagued by fragile interdependencies, or relying on restrictive, simplistic frameworks. Meanwhile, unconstrained, single-shot AI code generation promises speed but yields unstructured, unauditable chaos. The core challenge is combining the control and expressiveness of custom development with the efficiency of AI generation under strict constraints. To address this, we introduce BONSAI, a mixed-initiative workspace for the multi-agent co-development of VA applications. BONSAI utilizes a modular four-layer architecture (hardware, services, orchestration, application) that allows human and AI developers to independently contribute reusable components. The workspace incorporates this architecture into a structured four-phase development process (plan, design, monitor, and review), ensuring distributed agency and full provenance, where all human and AI contributions are structurally bounded and tracked. We evaluate BONSAI through case studies demonstrating the efficient creation of novel tools and the rapid reconstruction of complex VA applications directly from research paper descriptions. Ultimately, this paper contributes a conceptual workflow, a scalable architecture, and an integrated system that successfully balances AI's generative speed with the structural rigor required for complex VA development.
The FASER experiment is located in the Large Hadron Collider (LHC) complex at CERN, 480 m downstream of the ATLAS collision point and aligned with the beam-collision-axis. The experiment was designed to search for light, weakly-interacting new-particles which could be produced in the LHC collisions, and, for the first-time, to study high-energy neutrinos of all flavours originating at a particle collider. This review article presents the status of FASER up to early-2026. This includes details of the FASER detector design, operation, performance and physics results, as well as briefly mentioning upgrades that have been installed since the start of FASER. In addition, future plans for the experiment are detailed.
Gamma-ray bursts (GRBs) are generally classified as Type~I GRBs, which originate from compact binary mergers, and Type~II GRBs, which originate from massive collapsars. The traditional correspondence between short--Type~I GRBs and long--Type~II GRBs, separated by a duration of 2 seconds, has been challenged by recent observations of long GRBs associated with kilonovae (i.e., Type~I-L GRBs) and a short GRB associated with a supernova. In this paper, we focus on GRBs with precursor emission (PE) and compile 366 GRBs detected by Fermi/GBM. Applying the unsupervised machine learning methods t-SNE and UMAP, we are able to distinguish Type~I (including subclass Type~I-L) and Type~II GRBs for the first time and identify PE as a key feature for distinguishing GRBs of different origins. Inspired by results of machine learning, we propose a diagnostic parameter, the $E_{\rm p,ME}$-precursor index ($EPI$), defined as ${\rm log_{10}}(E_{\rm p,ME}^{2}/(T_{\rm 100,PE}T_{\rm 100,QE1}^{1/2}T_{\rm MVT,PE}))$, where most Type~I GRBs have $EPI > 6.2$ and most Type~II GRBs have $EPI < 6.2$. This parameter can help the community to diagnose the origin of any GRB with PE based solely on its prompt emission and rapidly plan for follow-up observations. The validation using Swift GRBs provides illustrative evidence that our method may also be applicable to GRBs observed by instruments other than Fermi.
We present a hybrid framework to support prognostics of the clogging degradation phenomenon in tube support plates for digital twins of steam generators in pressurized water reactors. The proposed approach combines a physics-based simulation code, heterogeneous and sparse observational data, and several uncertainty quantification techniques to obtain a robust estimate of the steam generator remaining useful life associated with the clogging rate. The proposed framework is compatible with a digital twin platform to assist maintenance planning of EDF steam generators.
Efficient and robust path planning hinges on combining all accessible information sources. In particular, the task of path planning for robotic environmental exploration and monitoring depends highly on the current belief of the world. To capture the uncertainty in the belief, we present a Gaussian process based path planning method that adapts to multi-modal environmental sensing data and incorporates state and input constraints. To solve the path planning problem, we optimize over future waypoints in a receding horizon fashion, and our cost is thus a function of the Gaussian process posterior over all these waypoints. We demonstrate this method, dubbed OLAhGP, on an autonomous surface vessel using oceanic algal bloom data from both a high-fidelity model and in-situ sensing data in a monitoring scenario. Our simulated and experimental results demonstrate significant improvement over existing methods. With the same number of samples, our method generates more informative paths and achieves greater accuracy in identifying algal blooms in chlorophyll a rich waters, measured with respect to total misclassification probability and binary misclassification rate over the domain of interest.
Vision-Language Models (VLMs) have become central to autonomous driving systems, yet their deployment is severely bottlenecked by the massive computational overhead of multi-view camera and multi-frame video input. Existing token pruning methods, primarily designed for single-image inputs, treat each frame or view in isolation and thus fail to exploit the inherent spatio-temporal redundancies in driving scenarios. To bridge this gap, we propose ST-Prune, a training-free, plug-and-play framework comprising two complementary modules: Motion-aware Temporal Pruning (MTP) and Ring-view Spatial Pruning (RSP). MTP addresses temporal redundancy by encoding motion volatility and temporal recency as soft constraints within the diversity selection objective, prioritizing dynamic trajectories and current-frame content over static historical background. RSP further resolves spatial redundancy by exploiting the ring-view camera geometry to penalize bilateral cross-view similarity, eliminating duplicate projections and residual background that temporal pruning alone cannot suppress. These two modules together constitute a complete spatio-temporal pruning process, preserving key scene information under strict compression. Validated across four benchmarks spanning perception, prediction, and planning, ST-Prune establishes new state-of-the-art for training-free token pruning. Notably, even at 90\% token reduction, ST-Prune achieves near-lossless performance with certain metrics surpassing the full-model baseline, while maintaining inference speeds comparable to existing pruning approaches.
Language-guided unmanned aerial vehicles (UAVs) often fail not from bad reasoning or perception, but from execution mismatch: the gap between a planned trajectory and the controller's ability to track it when the real dynamics differ from training (mass changes, drag shifts, actuator delay, wind). We propose AeroBridge-TTA, a language-conditioned control pipeline that targets this gap with test-time adaptation. It has three parts: a language encoder that maps the command into a subgoal, an adaptive policy conditioned on the subgoal and a learned latent, and a test-time adaptation (TTA) module that updates the latent online from observed transitions. On five language-conditioned UAV tasks under 13 mismatch conditions with the same domain randomization, AeroBridge-TTA ties a strong PPO-MLP baseline in-distribution and wins all 5 out-of-distribution (OOD) conditions, +22.0 pts on average (62.7% vs. 40.7%); the +8.5 pt overall gain comes entirely from the OOD regime. A same-weights ablation that only changes the step size $α$ shows the latent update itself is responsible for a $4.6\times$ OOD lift.
Large-scale cloud security platforms must continuously query millions of structured cloud resource records distributed across thousands of tenant accounts. Broad, account-spanning queries saturate database infrastructure, producing P95 latencies exceeding 60 seconds. We identify buffer cache pressure as the dominant latency driver: in a controlled experiment, the same query executing with the same plan completed in 3.7 seconds when its working set was memory-resident and 94 seconds when concurrent load had evicted those pages. No query plan optimization can address this; the only effective intervention is reducing the number of pages each query must touch. We present the Heuristic Search Space Partitioning System (HSSPS), a query-time optimization layer that logically partitions the search space through dynamic predicate injection, without schema modification. A two-phase heuristic engine selects partition key values and scores candidate query plans before execution. A client-side page token maintains cross-partition traversal state without server-side sessions, enabling horizontal scalability. Controlled evaluation across representative query types demonstrates 50-97% P95 latency reduction (95-97% on high-cardinality queries), 8-10x throughput improvement, and 41x reduction in average active sessions. Production rollout across live multi-tenant traffic reduced P95 latency from 61s to 2s across successive releases, sustained over 14,000 eligible queries per measurement window. The technique generalizes to any multi-tenant system where broad queries execute against large shared databases and physical schema modification is impractical.
Efficient construction of models capturing the preconditions and effects of actions is essential for applying AI planning in real-world domains. Extensive prior work has explored learning such models from high-level descriptions of state and/or action sequences. In this paper, we tackle a more challenging setting: learning lifted action models from sequences of state images, without action observation. We propose a deep learning framework that jointly learns state prediction, action prediction, and a lifted action model. We also introduce a mixed-integer linear program (MILP) to prevent prediction collapse and self-reinforcing errors among predictions. The MILP takes the predicted states, actions, and action model over a subset of traces and solves for logically consistent states, actions, and action model that are as close as possible to the original predictions. Pseudo-labels extracted from the MILP solution are then used to guide further training. Experiments across multiple domains show that integrating MILP-based correction helps the model escape local optima and converge toward globally consistent solutions.
Understanding the seeing conditions is crucial for astronomical observations using a ground-based telescope. This study analyzes long-term atmospheric data (2002-2021) from the ERA5 dataset to assess the seeing conditions at the new Timau National Observatory in Indonesia, which will house a 3.8-meter optical telescope. While the ERA5 dataset shows good agreement with radiosonde data for temperature and wind speed, it tends to underestimate seeing at Eltari Airport, Kupang. Despite this discrepancy, the ERA5 data suggest a median seeing of 0.79 arcseconds at Timau, with optimal seeing conditions in March and December and greater variability during the May to September dry season. These findings are crucial for the planning and operation of the observatory, which requires excellent seeing conditions for its three-band optical imager and a near-infrared camera. Although the seeing at Timau is not as good as some other observatories, the conditions at Timau make it an observatory that has good prospects for equatorial regions.
Multimodal empathetic response generation (MERG) aims to generate emotionally engaging and empathetic responses based on users' multimodal contexts. Existing approaches usually rely on an implicit one-pass generation paradigm from multimodal context to the final response, which overlooks two intrinsic characteristics of MERG: (1) Human perception of emotional cues is inherently structured rather than a direct mapping. The conventional paradigm neglects the hierarchical progression of emotion perception, leading to distorted emotional judgments. (2) Given the inherent complexity and ambiguity of human emotions, the conventional paradigm is prone to significant emotional biases, ultimately resulting in suboptimal empathy. In this paper, we propose a multi-agent framework for MERG, which enhances empathy through structured reasoning and reflective refinement. Specifically, we first introduce a structured empathetic reasoning-to-generation module that explicitly decomposes response generation via multimodal perception, consistency-aware emotion forecasting, pragmatic strategy planning, and strategy-guided response generation, providing a clearer intermediate path from multimodal evidence to response realization. Besides, we develop a global reflection and refinement module, in which a global reflection agent performs step-wise auditing over intermediate states and the generated response, eliminating existing emotional biases and empathy errors, and triggering targeted regeneration. Overall, such a closed-loop framework enables our model to gradually improve the accuracy of emotion perception and eliminate emotion biases during the iteration process. Experiments on several benchmarks, e.g., IEMOCAP and MELD, demonstrate that our model has superior empathic response generation capabilities compared to state-of-the-art methods.