We identify three critical gaps that hinder the na ıve integration of SD into RL systems: diminishing speedups at large batch sizes, drafter staleness under continual actor updates, and drafter-induced policy degradation. Among these stages, generation is consistently the dominant bottleneck (Zhong et al., 2025a). Classic policy optimization methods such as PPO (Schul-man et al., 2017; 2015) combine trajectory-level rewards A natural optimization to address this bottleneck is speculative decoding (SD) (Leviathan et al., 2023; Chen et al., SD has already been widely adopted in LLM serving systems (e.g., SGLang (Zheng et al., 2024), Among various SD variants, EAGLE-3 (Li et al., 2025) represents the current state of the art, achieving the Overview of SD in RL training and our proposed ReSpec system. Consequently, a single static SD configuration cannot provide reliable speedups across diverse RL workloads, as shown in Figure 3. (G2) Drafter staleness under continual actor updates. As the actor (i.e., target) model evolves with each policy update, a fixed drafter rapidly becomes misaligned with the actor Moreover, this variance increases with drafter staleness (G2), causing a higher ratio of impoverished trajectories.

Traditional collaborative filtering (CF) methods, relying on user - item interaction matrices, effectively capture latent patterns but face challenges such as data sparsity, cold - start problems, and limited contextual integration . To address these issues, M atrix F actorization (MF) techniques such as Singular Value Decomposition (SVD) and Alternating Least Squares (ALS) [1, 2 ] have been employed, improving accuracy but still struggling with sparsity and contextual richness. This has spurred the integration of side information, such as item content, social networks, and knowledge graphs, to enhance CF performance [3, 4 ] . Graph - based CF methods have emerged as a promising alternative, leveraging graph structures to model user - item interactions more effectively. Early approaches, such as ItemRank [5] and BiRank [6], used label propagation but lacked optimization capabilities . More advanced techniques, like HOP - Rec [7], integrated random walks with BPR . However, these models remain highly sensitive to hyperparameter tuning and often fail to capture high - order collaborative signals effectively . Graph Neural Networks (GNNs) have revolutionized recommendation systems by capturing complex user - item interactions, particularly in sparse data scenarios . Models like GC - MC [8] and PinSage [9] enhance user - item and item - item relationships, while SpectralCF [10] leverages spectral convolutions but faces scalability challenges.

Identifying and addressing security issues during the early phase of the development lifecycle is critical for mitigating the long-term negative impacts on software systems. Code review serves as an effective practice that enables developers to check their teammates' code before integration into the codebase. To streamline the generation of review comments, various automated code review approaches have been proposed, where LLM-based methods have significantly advanced the capabilities of automated review generation. However, existing models primarily focus on general-purpose code review, their effectiveness in identifying and addressing security-related issues remains underexplored. Moreover, adapting existing code review approaches to target security issues faces substantial challenges, including data scarcity and inadequate evaluation metrics. To address these limitations, we propose SecureReviewer, a new approach designed for enhancing LLMs' ability to identify and resolve security-related issues during code review. Specifically, we first construct a dataset tailored for training and evaluating secure code review capabilities. Leveraging this dataset, we fine-tune LLMs to generate code review comments that can effectively identify security issues and provide fix suggestions with our proposed secure-aware fine-tuning strategy. To mitigate hallucination in LLMs and enhance the reliability of their outputs, we integrate the RAG technique, which grounds the generated comments in domain-specific security knowledge. Additionally, we introduce SecureBLEU, a new evaluation metric designed to assess the effectiveness of review comments in addressing security issues. Experimental results demonstrate that SecureReviewer outperforms state-of-the-art baselines in both security issue detection accuracy and the overall quality and practical utility of generated review comments.

With the rapid development of large language models (LLMs), various LLM-based works have been widely applied in educational fields. However, most existing LLMs and their benchmarks focus primarily on the knowledge dimension, largely neglecting the evaluation of cultivation capabilities that are essential for real-world educational scenarios. Additionally, current benchmarks are often limited to a single subject or question type, lacking sufficient diversity. This issue is particularly prominent within the Chinese context. To address this gap, we introduce OmniEduBench, a comprehensive Chinese educational benchmark. OmniEduBench consists of 24.602K high-quality question-answer pairs. The data is meticulously divided into two core dimensions: the knowledge dimension and the cultivation dimension, which contain 18.121K and 6.481K entries, respectively. Each dimension is further subdivided into 6 fine-grained categories, covering a total of 61 different subjects (41 in the knowledge and 20 in the cultivation). Furthermore, the dataset features a rich variety of question formats, including 11 common exam question types, providing a solid foundation for comprehensively evaluating LLMs' capabilities in education. Extensive experiments on 11 mainstream open-source and closed-source LLMs reveal a clear performance gap. In the knowledge dimension, only Gemini-2.5 Pro surpassed 60\% accuracy, while in the cultivation dimension, the best-performing model, QWQ, still trailed human intelligence by nearly 30\%. These results highlight the substantial room for improvement and underscore the challenges of applying LLMs in education.

The rapid advancement of deep neural networks (DNNs) heavily relies on large-scale, high-quality datasets. However, unauthorized commercial use of these datasets severely violates the intellectual property rights of dataset owners. Existing backdoor-based dataset ownership verification methods suffer from inherent limitations: poison-label watermarks are easily detectable due to label inconsistencies, while clean-label watermarks face high technical complexity and failure on high-resolution images. Moreover, both approaches employ static watermark patterns that are vulnerable to detection and removal. To address these issues, this paper proposes a sample-specific clean-label backdoor watermarking (i.e., SSCL-BW). By training a U-Net-based watermarked sample generator, this method generates unique watermarks for each sample, fundamentally overcoming the vulnerability of static watermark patterns. The core innovation lies in designing a composite loss function with three components: target sample loss ensures watermark effectiveness, non-target sample loss guarantees trigger reliability, and perceptual similarity loss maintains visual imperceptibility. During ownership verification, black-box testing is employed to check whether suspicious models exhibit predefined backdoor behaviors. Extensive experiments on benchmark datasets demonstrate the effectiveness of the proposed method and its robustness against potential watermark removal attacks.

Artificial intelligence in healthcare requires models that are accurate and interpretable. We advance mechanistic interpretability in medical vision by applying Medical Sparse Autoencoders (MedSAEs) to the latent space of MedCLIP, a vision-language model trained on chest radiographs and reports. To quantify interpretability, we propose an evaluation framework that combines correlation metrics, entropy analyzes, and automated neuron naming via the MedGEMMA foundation model. Experiments on the CheXpert dataset show that MedSAE neurons achieve higher monosemanticity and interpretability than raw MedCLIP features. Our findings bridge high-performing medical AI and transparency, offering a scalable step toward clinically reliable representations.

The emergence of agentic Artificial Intelligence (AI) is set to trigger a "Cambrian explosion" of new kinds of personhood. This paper proposes a pragmatic framework for navigating this diversification by treating personhood not as a metaphysical property to be discovered, but as a flexible bundle of obligations (rights and responsibilities) that societies confer upon entities for a variety of reasons, especially to solve concrete governance problems. We argue that this traditional bundle can be unbundled, creating bespoke solutions for different contexts. This will allow for the creation of practical tools -- such as facilitating AI contracting by creating a target "individual" that can be sanctioned -- without needing to resolve intractable debates about an AI's consciousness or rationality. We explore how individuals fit in to social roles and discuss the use of decentralized digital identity technology, examining both "personhood as a problem", where design choices can create "dark patterns" that exploit human social heuristics, and "personhood as a solution", where conferring a bundle of obligations is necessary to ensure accountability or prevent conflict. By rejecting foundationalist quests for a single, essential definition of personhood, this paper offers a more pragmatic and flexible way to think about integrating AI agents into our society.

The prohibitive cost of evaluating large language models (LLMs) on comprehensive benchmarks necessitates the creation of small yet representative data subsets (i.e., tiny benchmarks) that enable efficient assessment while retaining predictive fidelity. Current methods for this task operate under a model-centric paradigm, selecting benchmarking items based on the collective performance of existing models. Such approaches are limited by large upfront costs, an inability to immediately handle new benchmarks (`cold-start'), and the fragile assumption that future models will share the failure patterns of their predecessors. In this work, we challenge this paradigm and propose a item-centric approach to benchmark subset selection, arguing that selection should be based on the intrinsic properties of the task items themselves, rather than on model-specific failure patterns. We instantiate this item-centric efficient benchmarking approach via a novel method, Scales++, where data selection is based on the cognitive demands of the benchmark samples. Empirically, we show Scales++ reduces the upfront selection cost by over 18x while achieving competitive predictive fidelity. On the Open LLM Leaderboard, using just a 0.5\% data subset, we predict full benchmark scores with a 2.9% mean absolute error. We demonstrate that this item-centric approach enables more efficient model evaluation without significant fidelity degradation, while also providing better cold-start performance and more interpretable benchmarking.

Autonomous robots in orchards require real-time 3D scene understanding despite repetitive row geometry, seasonal appearance changes, and wind-driven foliage motion. We present AgriGS-SLAM, a Visual--LiDAR SLAM framework that couples direct LiDAR odometry and loop closures with multi-camera 3D Gaussian Splatting (3DGS) rendering. Batch rasterization across complementary viewpoints recovers orchard structure under occlusions, while a unified gradient-driven map lifecycle executed between keyframes preserves fine details and bounds memory. Pose refinement is guided by a probabilistic LiDAR-based depth consistency term, back-propagated through the camera projection to tighten geometry-appearance coupling. We deploy the system on a field platform in apple and pear orchards across dormancy, flowering, and harvesting, using a standardized trajectory protocol that evaluates both training-view and novel-view synthesis to reduce 3DGS overfitting in evaluation. Across seasons and sites, AgriGS-SLAM delivers sharper, more stable reconstructions and steadier trajectories than recent state-of-the-art 3DGS-SLAM baselines while maintaining real-time performance on-tractor. While demonstrated in orchard monitoring, the approach can be applied to other outdoor domains requiring robust multimodal perception.

Millions of people take surveys every day, from market polls and academic studies to medical questionnaires and customer feedback forms. These datasets capture valuable insights, but their scale and structure present a unique challenge for large language models (LLMs), which otherwise excel at few-shot reasoning over open-ended text. Yet, their ability to process questionnaire data or lists of questions crossed with hundreds of respondent rows remains underexplored. Current retrieval and survey analysis tools (e.g., Qualtrics, SPSS, REDCap) are typically designed for humans in the workflow, limiting such data integration with LLM and AI-empowered automation. This gap leaves scientists, surveyors, and everyday users without evidence-based guidance on how to best represent questionnaires for LLM consumption. We address this by introducing QASU (Questionnaire Analysis and Structural Understanding), a benchmark that probes six structural skills, including answer lookup, respondent count, and multi-hop inference, across six serialization formats and multiple prompt strategies. Experiments on contemporary LLMs show that choosing an effective format and prompt combination can improve accuracy by up to 8.8% points compared to suboptimal formats. For specific tasks, carefully adding a lightweight structural hint through self-augmented prompting can yield further improvements of 3-4% points on average. By systematically isolating format and prompting effects, our open source benchmark offers a simple yet versatile foundation for advancing both research and real-world practice in LLM-based questionnaire analysis.