In a closed-door simulation, insurers played out their response to a mass disruption by China's Volt Typhoon hackers--and found a nightmare scenario. It's around an hour and 10 minutes into the role-playing game I've been invited to observe, a simulated catastrophic cyberattack on US water utilities, when the whole thing begins to feel less like a fun afternoon playing Dungeons & Dragons and more like a plausible threat to civilization. A full 24 hours of in-game time have passed since hackers disrupted 5,000 water utilities across the United States in this imagined scenario. Joshua Corman, the former Cybersecurity and Infrastructure Security Agency strategist serving as our dungeon master, stands at the front of a conference space in an office tower high above Times Square, narrating the latest updates to the game's participants, a few dozen insurance executives set up in six teams. All of them have gone disturbingly silent. It's about to get harder," Corman says. "I'm going to share a few things, and it's going to hurt." It is, of course, still the same April afternoon as when we started--but in game time, the second-order effects of widespread water outages have started to become clear. Food refrigeration systems are failing at cold storage warehouses. Water-dependent drug and chemical manufacturing has been bottlenecked, leading to insulin shortages. Data centers' cooling systems are failing, causing outages of cloud services. Most critically, 2,000 hospitals are without water, hampering patient care and in some cases leading to evacuations as HVAC systems shut down and the July heat--the game takes place just before Independence Day in 2027--bakes facilities. Worse yet, Corman is playing a looping video onscreen, at the front of the room, showing a burst water main: The hackers have managed to trigger not just IT disruption but also, in at least some cases, real physical destruction that will take far longer to fix. "Everyone downstream is without water pressure," Corman says. "There are no breaks in real incident response," Corman explains just before the giant water pipe starts gushing onscreen. "If you have to go to the bathroom, go to the bathroom.

Despite the rapid advancement of LLM-based agents, the reliable evaluation of their safety and security remains a significant challenge. Existing rule-based or LLM-based evaluators often miss dangers in agents' step-by-step actions, overlook subtle meanings, fail to see how small issues compound, and get confused by unclear safety or security rules. To overcome this evaluation crisis, we introduce AgentAuditor, a universal, training-free, memory-augmented reasoning framework that empowers LLM evaluators to emulate human expert evaluators. AgentAuditor constructs an experiential memory by having an LLM adaptively extract structured semantic features (e.g., scenario, risk, behavior) and generate associated chain-of-thought reasoning traces for past interactions. A multi-stage, context-aware retrieval-augmented generation process then dynamically retrieves the most relevant reasoning experiences to guide the LLM evaluator's assessment of new cases. Moreover, we developed ASSEBench, the first benchmark designed to check how well LLM-based evaluators can spot both safety risks and security threats. ASSEBench comprises 2293 meticulously annotated interaction records, covering 15 risk types across 29 application scenarios. A key feature of ASSEBench is its nuanced approach to ambiguous risk situations, employing Strict and Lenient judgment standards. Experiments demonstrate that AgentAuditor not only consistently improves the evaluation performance of LLMs across all benchmarks but also sets a new state-of-the-art in LLM-as-a-judge for agent safety and security, achieving human-level accuracy.

As the era of autonomous agents making decisions on behalf of users unfolds, ensuring contextual integrity (CI) - what is the appropriate information to share while carrying out a certain task - becomes a central question to the field. We posit that CI demands a form of reasoning where the agent needs to reason about the context in which it is operating. To test this, we first prompt LLMs to reason explicitly about CI when deciding what information to disclose. We then extend this approach by developing a reinforcement learning (RL) framework that further instills in models the reasoning necessary to achieve CI. Using a synthetic, automatically created, dataset of only 700 examples but with diverse contexts and information disclosure norms, we show that our method substantially reduces inappropriate information disclosure while maintaining task performance across multiple model sizes and families. Importantly, improvements transfer from this synthetic dataset to established CI benchmarks such as PrivacyLens that has human annotations and evaluates privacy leakage of AI assistants in actions and tool calls. Our code is available at: https://github.com/EricGLan/CI-RL

Steering cooperative multi-agent reinforcement learning (MARL) towards desired outcomes is challenging, particularly when the global guidance from a human on the whole multi-agent system is impractical in a large-scale MARL. On the other hand, designing external mechanisms (e.g., intrinsic rewards and human feedback) to coordinate agents mostly relies on empirical studies, lacking a easy-to-use research tool. In this work, we employ multi-agent influence diagrams (MAIDs) as a graphical framework to address the above issues. First, we introduce the concept of MARL interaction paradigms (orthogonal to MARL learning paradigms), using MAIDs to analyze and visualize both unguided self-organization and global guidance mechanisms in MARL. Then, we design a new MARL interaction paradigm, referred to as the targeted intervention paradigm that is applied to only a single targeted agent, so the problem of global guidance can be mitigated. In implementation, we introduce a causal inference technique--referred to as Pre-Strategy Intervention (PSI)--to realize the targeted intervention paradigm. Since MAIDs can be regarded as a special class of causal diagrams, a composite desired outcome that integrates the primary task goal and an additional desired outcome can be achieved by maximizing the corresponding causal effect through the PSI. Moreover, the bundled relevance graph analysis of MAIDs provides a tool to identify whether an MARL learning paradigm is workable under the design of an MARL interaction paradigm. In experiments, we demonstrate the effectiveness of our proposed targeted intervention, and verify the result of relevance graph analysis.

We engage in the relatively underexplored task named thermal infrared image enhancement. Existing infrared image enhancement methods primarily focus on tackling individual degradations, such as noise, contrast, and blurring, making it difficult to handle coupled degradations. Meanwhile, all-in-one enhancement methods, commonly applied to RGB sensors, often demonstrate limited effectiveness due to the significant differences in imaging models. In sight of this, we first revisit the imaging mechanism and introduce a Progressive Prompt Fusion Network (PPFN). Specifically, the PPFN initially establishes prompt pairs based on the thermal imaging process. For each type of degradation, we fuse the corresponding prompt pairs to modulate the model's features, providing adaptive guidance that enables the model to better address specific degradations under single or multiple conditions. In addition, a Selective Progressive Training (SPT) mechanism is introduced to gradually refine the model's handling of composite cases to align the enhancement process, which not only allows the model to remove camera noise and retain key structural details, but also enhancing the overall contrast of the thermal image. Furthermore, we introduce the most high-quality, multi-scenarios infrared benchmark covering a wide range of scenarios. Extensive experiments substantiate that our approach not only delivers promising visual results under specific degradation but also significantly improves performance on complex degradation scenes, achieving a notable 8.76% improvement.

Image quality assessment (IQA) aims to quantify the visual quality of digital images consistent with human perceptual judgments. Commonly, IQA models are classified into full-reference (FR) and noreference (NR) approaches [47], depending on the availability of pristine-quality reference images. In this paper, we focus on NR-IQA due to its practical relevance in real-world scenarios where reference images are unavailable. Over the decades, NR-IQA has evolved from knowledge-driven [33, 12] to data-driven approaches [30, 19, 54], and shifted from regression-based to ranking-based [58, 59] techniques. Nevertheless, achieving strong model generalization (e.g., generalization to unseen image distortions) remains a significant, unresolved challenge, driving recent research toward multi-dataset training [6], active fine-tuning [44], and continual model adaptation [57]. The rapid advancement of vision-language models (VLMs) offers promising avenues for enhancing NR-IQA generalization by contextualizing it into broader vision tasks [51]. VLMs can effectively integrate multi-modal information, enabling understanding of both low-level image distortions (e.g., noise and blur) and high-level perceptual attributes (e.g., aesthetics and content semantics). This multi-modal semantic contextualization allows VLMs to articulate nuanced quality descriptions with stronger generalization. However, current NR-IQA methods mainly leverage VLMs through supervised fine-tuning (SFT), which face several critical limitations [49, 56].

Multimodal Large Language Models (MLLMs) have achieved considerable accuracy in Optical Character Recognition (OCR) from static images. However, their efficacy in video OCR is significantly diminished due to factors such as motion blur, temporal variations, and visual effects inherent in video content. To provide clearer guidance for training practical MLLMs, we introduce MMEVideoOCR benchmark, which encompasses a comprehensive range of video OCR application scenarios.

Human cognition is theorized to operate in two modes: fast, intuitive System 1 thinking and slow, deliberate System 2 thinking. While current Large Reasoning Models (LRMs) excel at System 2 thinking, their inability to perform fast thinking leads to high computational overhead and latency. In this work, we enable LRMs to approximate human intelligence through dynamic thinking speed adjustment, optimizing accuracy-efficiency trade-offs. Our approach addresses two key questions: (1) how to control thinking speed in LRMs, and (2) when to adjust it for optimal performance. For the first question, we identify the steering vector that governs slow-fast thinking transitions in LRMs' representation space.

End-to-end learning has emerged as a transformative paradigm for autonomous driving. However, the inherently multimodal nature of driving behaviors remains a fundamental challenge to robust deployment. We propose DiffE2E, a diffusionbased end-to-end autonomous driving framework. The architecture first performs multi-scale alignment of perception features from multiple sensors via a hierarchical bidirectional cross-attention mechanism.

Causal Representation Learning (CRL) aims to uncover the data-generating process and identify the underlying causal variables and relations, whose evaluation remains inherently challenging due to the requirement of known ground-truth causal variables and causal structure. Existing evaluations often rely on either simplistic synthetic datasets or downstream performance on real-world tasks, generally suffering a dilemma between realism and evaluative precision. In this paper, we introduce a new benchmark for CRL using high-fidelity simulated visual data that retains both realistic visual complexity and, more importantly, access to groundtruth causal generating processes. The dataset comprises around 200 thousand images and 3 million video frames across 24 sub-scenes in four domains: static image generation, dynamic physical simulations, robotic manipulations, and traffic situation analysis. These scenarios range from static to dynamic settings, simple to complex structures, and single to multi-agent interactions, offering a comprehensive testbed that hopefully bridges the gap between rigorous evaluation and real-world applicability. In addition, we provide flexible access to the underlying causal structures, allowing users to modify or configure them to align with the required assumptions in CRL, such as available domain labels, temporal dependencies, or intervention histories. Leveraging this benchmark, we evaluated representative CRL methods across diverse paradigms and offered empirical insights to assist practitioners and newcomers in choosing or extending appropriate CRL frameworks to properly address specific types of real problems that can benefit from the CRL perspective. Welcome to visit our: Project page: causal-verse.github.io,