Mask Diffusion-based Vision Language Models (MDVLMs) have achieved remarkable progress in multimodal understanding tasks. However, these models are unable to correct errors in generated tokens, meaning they lack self-correction capability. In this paper, we propose Recursive Introspection Mask Diffusion Vision Language Model (RIV), which equips the model with self-correction ability through two novel mechanisms. The first is Introspection Training, where an Introspection Model is introduced to identify errors within generated sequences. Introspection Training enables the model to detect not only grammatical and spelling mistakes, but more importantly, logical errors. The second is Recursive Inference. Beginning with the standard unmasking step, the learned Introspection Model helps to identify errors in the output sequence and remask them. This alternating ($\text{unmask}\rightarrow\text{introspection}\rightarrow\text{remask}$) process is repeated recursively until reliable results are obtained. Experimental results on multiple benchmarks demonstrate that the proposed RIV achieves state-of-the-art performance, outperforming most existing MDVLMs.

Structural changes in retinal blood vessels are critical biomarkers for the onset and progression of glaucoma and other ocular diseases. However, current vessel segmentation approaches largely rely on supervised learning and extensive manual annotations, which are costly, error-prone, and difficult to obtain in optical coherence tomography angiography. Here we present BioVessel-Net, an unsupervised generative framework that integrates vessel biostatistics with adversarial refinement and a radius-guided segmentation strategy. Unlike pixel-based methods, BioVessel-Net directly models vascular structures with biostatistical coherence, achieving accurate and explainable vessel extraction without labeled data or high-performance computing. To support training and evaluation, we introduce RetinaMix, a new benchmark dataset of 2D and 3D OCTA images with high-resolution vessel details from diverse populations. Experimental results demonstrate that BioVessel-Net achieves near-perfect segmentation accuracy across RetinaMix and existing datasets, substantially outperforming state-of-the-art supervised and semi-supervised methods. Together, BioVessel-Net and RetinaMix provide a label-free, computationally efficient, and clinically interpretable solution for retinal vessel analysis, with broad potential for glaucoma monitoring, blood flow modeling, and progression prediction. Code and dataset are available: https://github.com/VikiXie/SatMar8.

Effective guardrails are essential for safely deploying LLM-based agents in critical applications. Despite recent advances, existing guardrails suffer from two fundamental limitations: (i) they apply uniform guardrail policies to all users, ignoring that the same agent behavior can harm some users while being safe for others; (ii) they check each response in isolation, missing how risks evolve and accumulate across multiple interactions. To solve these issues, we propose PSG-Agent, a personalized and dynamic system for LLM-based agents. First, PSG-Agent creates personalized guardrails by mining the interaction history for stable traits and capturing real-time states from current queries, generating user-specific risk thresholds and protection strategies. Second, PSG-Agent implements continuous monitoring across the agent pipeline with specialized guards, including Plan Monitor, Tool Firewall, Response Guard, Memory Guardian, that track cross-turn risk accumulation and issue verifiable verdicts. Finally, we validate PSG-Agent in multiple scenarios including healthcare, finance, and daily life automation scenarios with diverse user profiles. It significantly outperform existing agent guardrails including LlamaGuard3 and AGrail, providing an executable and auditable path toward personalized safety for LLM-based agents.

Simulated data-assisted SAR target recognition methods are the research hotspot currently, devoted to solving the problem of limited samples. Existing works revolve around simulated images, but the large amount of irrelevant information embedded in the images, such as background, noise, etc., seriously affects the quality of the migrated information. Our work explores a new simulated data to migrate purer and key target knowledge, i.e., forward scattering center model (FSCM) which models the actual local structure of the target with strong physical meaning and interpretability. To achieve this purpose, multi-level heterogeneous knowledge transfer (MHKT) network is proposed, which fully migrates FSCM knowledge from the feature, distribution and category levels, respectively. Specifically, we permit the more suitable feature representations for the heterogeneous data and separate non-informative knowledge by task-associated information selector (TAIS), to complete purer target feature migration. In the distribution alignment, the new metric function maximum discrimination divergence (MDD) in target generic knowledge transfer (TGKT) module perceives transferable knowledge efficiently while preserving discriminative structure about classes. Moreover, category relation knowledge transfer (CRKT) module leverages the category relation consistency constraint to break the dilemma of optimization bias towards simulation data due to imbalance between simulated and measured data. Such stepwise knowledge selection and migration will ensure the integrity of the migrated FSCM knowledge. Notably, extensive experiments on two new datasets formed by FSCM data and measured SAR images demonstrate the superior performance of our method.

Post-training, which elicits a pretrained Base model into the corresponding Instruct model, is widely considered to be superficial. In this work, we first reinforce this hypothesis by providing novel quantitative evidence from the weight level that the effective rank (eRank) remains negligibly changed. However, this superficiality also suffers a critical trade-off, improving the exploitation capabilities at the cost of limiting its exploration. To tackle this issue, we propose Timber, a simple yet effective training-free method that enhances the exploration capability of the Instruct model while preserving its exploitation. The key insight is to partially revert Instruct towards the paired Base model by subtle yet targeted refinement of the weight deltas. Extensive experiments on Llama and Qwen series demonstrate that Timber consistently improves vanilla Instruct models, particularly on Pass@k performance. Our findings offer new insights into the post-training stage at the weight level and practical strategies to refine the Instruct model without training. Large Language Models (LLMs), such as Qwen3 (Y ang et al., 2025), Llama 3 (Grattafiori et al., 2024), and Deepseek R1 (Guo et al., 2025), have achieved superior success in Natural Language Process (NLP), especially in reasoning tasks (Huang & Chang, 2022). To train these LLMs, a Base model is first pretrained on huge amounts of data. After that, a post-training stage is applied to train an Instruct model, adapting supervised finetuning (SFT) and reinforcement learning (RL) to elicit alignment and reasoning ability (Y ang et al., 2025). The post-training stage tends to be superficial, i.e., post-training only utilizes the pattern contained in the Base model acquired during pre-training (Y ue et al., 2025; Zhou et al., 2023a; Y e et al., 2025; Muennighoff et al., 2025). In this paper, we investigate the Base and Instruct models through the lens of effective rank (eRank, (Roy & V etterli, 2007)), providing a novel weight-level perspective on the superficiality of post-training. As shown in Figure 1, the eRanks of corresponding linear layers from the Base and Instruct models are almost identical. We can find that post-training induces only negligible changes to the effective dimensionality, offering new supporting evidence from the weight level for its superficiality.

We present a holistic framework for continual model merging that intervenes at three critical stages: pre-merging, during merging, and post-merging-to address two fundamental challenges in continual learning. In particular, conventional approaches either maintain a growing list of per-domain task vectors, leading to scalability issues or rely solely on weight-space merging when old data is inaccessible, thereby losing crucial functional information. Our method overcomes these limitations by first fine-tuning the main model within its tangent space on domain-specific data; this linearization amplifies per-task weight disentanglement, effectively mitigating across-task interference. During merging, we leverage functional information from available optimizer states beyond mere parameter averages to avoid the need to revisit old data. Finally, a post-merging correction aligns the representation discrepancy between pre- and post-merged models, reducing bias and enhancing overall performance-all while operating under constant memory constraints without accessing historical data. Extensive experiments on standard class-incremental and domain-incremental benchmarks demonstrate that our approach not only achieves competitive performance but also provides a scalable and efficient solution to the catastrophic forgetting problem.

Multi-turn agent systems based on Large Language Models (LLMs) have been increasingly popular for software engineering tasks. While LLM agents show decent effectiveness, the high computational cost of input tokens due to the ever-growing trajectory remains an efficiency concern for their applications. Efficiency is largely neglected in existing studies and agent products, and this paper fills the gap by introducing an inference-time trajectory reduction approach to reduce the cost of agents. Through analyzing existing agent trajectories, we demonstrate that useless, redundant, and expired information is widespread in all trajectories, which can be identified and reduced without harming the agent's performance. We then design a simple yet effective trajectory reduction approach, AgentDiet, which automatically removes such waste information. We implement AgentDiet on a top-performing coding agent, and the evaluation on two LLMs and two benchmarks shows that AgentDiet can reduce input tokens by 39.9% ~ 59.7%, or the final computational cost by 21.1% ~ 35.9%, while maintaining the same agent performance. This indicates that trajectory reduction is a promising direction for agent systems.

Enterprise teams rely on the Jira Query Language (JQL) to retrieve and filter issues from Jira. Yet, to our knowledge, there is no open, real-world, execution-based benchmark for mapping natural language queries to JQL. We introduce Jackal, a novel, large-scale text-to-JQL benchmark comprising 100,000 natural language (NL) requests paired with validated JQL queries and execution-based results on a live Jira instance with over 200,000 issues. To reflect real-world usage, each JQL query is associated with four types of user requests: (i) Long NL, (ii) Short NL, (iii) Semantically Similar, and (iv) Semantically Exact. We release Jackal, a corpus of 100,000 text-to-JQL pairs, together with an execution-based scoring toolkit, and a static snapshot of the evaluated Jira instance for reproducibility. We report text-to-JQL results on 23 Large Language Models (LLMs) spanning parameter sizes, open and closed source models, across execution accuracy, exact match, and canonical exact match. In this paper, we report results on Jackal-5K, a 5,000-pair subset of Jackal. On Jackal-5K, the best overall model (Gemini 2.5 Pro) achieves only 60.3% execution accuracy averaged equally across four user request types. Performance varies significantly across user request types: (i) Long NL (86.0%), (ii) Short NL (35.7%), (iii) Semantically Similar (22.7%), and (iv) Semantically Exact (99.3%). By benchmarking LLMs on their ability to produce correct and executable JQL queries, Jackal exposes the limitations of current state-of-the-art LLMs and sets a new, execution-based challenge for future research in Jira enterprise data.

Machine learning (ML) assets, such as models, datasets, and metadata, are central to modern ML workflows. Despite their explosive growth in practice, these assets are often underutilized due to fragmented documentation, siloed storage, inconsistent licensing, and lack of unified discovery mechanisms, making ML-asset management an urgent challenge. This tutorial offers a comprehensive overview of ML-asset management activities across its lifecycle, including curation, discovery, and utilization. We provide a categorization of ML assets, and major management issues, survey state-of-the-art techniques, and identify emerging opportunities at each stage. We further highlight system-level challenges related to scalability, lineage, and unified indexing. Through live demonstrations of systems, this tutorial equips both researchers and practitioners with actionable insights and practical tools for advancing ML-asset management in real-world and domain-specific settings.

Hierarchical coarse-to-fine policy, where a coarse branch predicts a region of interest to guide a fine-grained action predictor, has demonstrated significant potential in robotic 3D manipulation tasks by especially enhancing sample efficiency and enabling more precise manipulation. However, even augmented with pre-trained models, these hierarchical policies still suffer from generalization issues. To enhance generalization to novel instructions and environment variations, we propose Coarse-to-fine Language-Aligned manipulation Policy (CLAP), a framework that integrates three key components: 1) task decomposition, 2) VLM fine-tuning for 3D keypoint prediction, and 3) 3D-aware representation. Through comprehensive experiments in simulation and on a real robot, we demonstrate its superior generalization capability. Specifically, on GemBench, a benchmark designed for evaluating generalization, our approach achieves a 12% higher average success rate than the SOT A method while using only 1/5 of the training trajectories. In real-world experiments, our policy, trained on only 10 demonstrations, successfully generalizes to novel instructions and environments. Robot learning, especially via imitation learning, has demonstrated promising success in enabling robots to solve complex 3D manipulation tasks (Intelligence et al., 2025; Liu et al., 2024). However, scaling these methods to a broader range of real-world applications (e.g., industrial, service, or home robotics) requires enhancing both (G1) their generalization to environment variations, and (G2) their skill compositional generalization.