Procedural activity assistants potentially support humans in a variety of settings, from our daily lives, e.g., cooking or assembling flat-pack furniture, to professional situations, e.g., manufacturing or biological experiments. Despite its potential use cases, the system development tailored for such an assistant is still underexplored. In this paper, we propose a novel framework, called TAMA, a Tool-Augmented Multimodal Agent, for procedural activity understanding. TAMA enables interleaved multimodal reasoning by making use of multimedia-returning tools in a training-free setting. Our experimental result on the multimodal procedural QA dataset, ProMQA-Assembly, shows that our approach can improve the performance of vision-language models, especially GPT-5 and MiMo-VL. Furthermore, our ablation studies provide empirical support for the effectiveness of two features that characterize our framework, multimedia-returning tools and agentic flexible tool selection. We believe our proposed framework and experimental results facilitate the thinking with images paradigm for video and multimodal tasks, let alone the development of procedural activity assistants.

Financial fraud detection in real-world scenarios presents significant challenges due to the subtlety and dispersion of evidence across complex, multi-year financial disclosures. In this work, we introduce a novel multi-agent reasoning framework AuditAgent, enhanced with auditing domain expertise, for fine-grained evidence chain localization in financial fraud cases. Leveraging an expert-annotated dataset constructed from enforcement documents and financial reports released by the China Securities Regulatory Commission, our approach integrates subject-level risk priors, a hybrid retrieval strategy, and specialized agent modules to efficiently identify and aggregate cross-report evidence. Extensive experiments demonstrate that our method substantially outperforms General-Purpose Agent paradigm in both recall and interpretability, establishing a new benchmark for automated, transparent financial forensics. Our results highlight the value of domain-specific reasoning and dataset construction for advancing robust financial fraud detection in practical, real-world regulatory applications.

Despite rapid progress in autonomous robotics, executing complex or long-horizon tasks remains a fundamental challenge. Most current approaches follow an open-loop paradigm with limited reasoning and no feedback, resulting in poor robustness to environmental changes and severe error accumulation. We present RoboPilot, a dual-thinking closed-loop framework for robotic manipulation that supports adaptive reasoning for complex tasks in real-world dynamic environments. RoboPilot leverages primitive actions for structured task planning and flexible action generation, while introducing feedback to enable replanning from dynamic changes and execution errors. Chain-of-Thought reasoning further enhances high-level task planning and guides low-level action generation. The system dynamically switches between fast and slow thinking to balance efficiency and accuracy. To systematically evaluate the robustness of RoboPilot in diverse robot manipulation scenarios, we introduce RoboPilot-Bench, a benchmark spanning 21 tasks across 10 categories, including infeasible-task recognition and failure recovery. Experiments show that RoboPilot outperforms state-of-the-art baselines by 25.9\% in task success rate, and the real-world deployment on an industrial robot further demonstrates its robustness in real-world settings.

Sorbonne Universit e, CNRS, LIP6, Paris, France Abstract--AI models are often regarded as valuable intellectual property due to the high cost of their development, the competitive advantage they provide, and the proprietary techniques involved in their creation. As a result, AI model stealing attacks pose a serious concern for AI model providers. In this work, we present a novel attack targeting wireless devices equipped with AI hardware accelerators. The attack unfolds in two phases. In the first phase, the victim's device is compromised with a hardware T rojan (HT) designed to covertly leak model weights through a hidden communication channel, without the victim realizing it. In the second phase, the adversary uses a nearby wireless device to intercept the victim's transmission frames during normal operation and incrementally reconstruct the complete weight matrix. The proposed attack is agnostic to both the AI model architecture and the hardware accelerator used. Additionally, we analyze the impact of bit error rates on the reception and propose an error mitigation technique. The effectiveness of the attack is evaluated based on the accuracy of the reconstructed models with stolen weights and the time required to extract them. Finally, we explore potential defense mechanisms. I. Introduction AI models are regarded as valuable assets because their development demands significant investment in data collection, computational resources, and training time. They also offer a competitive edge, as model performance frequently distinguishes companies in the same industry. Furthermore, these models embody proprietary insights, including specialized feature engineering, architectural decisions, and unique training methodologies.

Machine unlearning is an emerging technique that removes the influence of a subset of training data (forget set) from a model without full retraining, with applications including privacy protection, content moderation, and model correction. The key challenge lies in ensuring that the model completely forgets the knowledge of the forget set without compromising its overall utility. Existing unlearning methods for large language models (LLMs) often utilize auxiliary language models, retain datasets, or even commercial AI services for effective unlearning and maintaining the model utility. However, dependence on these external resources is often impractical and could potentially introduce additional privacy risks. In this work, we propose direct token optimization (DTO), a novel self-contained unlearning approach for LLMs that directly optimizes the token level objectives and eliminates the need for external resources. Given a sequence to unlearn, we identify two categories of tokens: target tokens, which capture critical knowledge for unlearning, and the remaining non-target tokens, which are crucial for maintaining the model utility. The former are used to optimize the unlearning objective, while the latter serve to preserve the model's performance. The experimental results show that the proposed DTO achieves up to 16.8 improvement in forget quality on several benchmark datasets than the latest baselines while maintaining a comparable level of model utility. Machine unlearning aims to remove the effect of a subset of training data (referred to as the forget set) from a trained model (Cao & Y ang, 2015). The concept was introduced in response to data protection regulations such as General Data Protection Regulation (GDPR) (Mantelero, 2013), which established the'right to be forgotten'. A successfully unlearned model should fully eliminate the influence of forget set (unlearning efficacy), and preserve overall performance (model utility). Additionally, the unlearning algorithm should be more efficient than retraining (efficiency).

This study reinterprets a Monte Carlo simulation of students' perceived success with generative AI (GenAI) through a Kantian-axiomatic lens. Building on prior work, theme-level survey statistics Ease of Use and Learnability, System Efficiency and Learning Burden, and Perceived Complexity and Integration from a representative dataset are used to generate 10,000 synthetic scores per theme on the [1,5] Likert scale. The simulated outputs are evaluated against the axioms of dense linear order without endpoints (DLO): irreflexivity, transitivity, total comparability (connectedness), no endpoints (no greatest and no least; A4-A5), and density (A6). At the data level, the basic ordering axioms (A1-A3) are satisfied, whereas no-endpoints (A4-A5) and density (A6) fail as expected. Likert clipping introduces minimum and maximum observed values, and a finite, discretized sample need not contain a value strictly between any two distinct scores. These patterns are read not as methodological defects but as markers of an epistemological boundary. Following Kant and Friedman, the findings suggest that what simulations capture finite, quantized observations cannot instantiate the ideal properties of an unbounded, dense continuum. Such properties belong to constructive intuition rather than to finite sampling alone. A complementary visualization contrasts the empirical histogram with a sine-curve proxy to clarify this divide. The contribution is interpretive rather than data-expansive: it reframes an existing simulation as a probe of the synthetic a priori structure underlying students' perceptions, showing how formal order-theoretic coherence coexists with principled failures of endpoint-freeness and density in finite empirical models.

Large language models (LLMs) have been extensively used for legal judgment prediction tasks based on case reports and crime history. However, with a surge in the availability of large vision language models (VLMs), legal judgment prediction systems can now be made to leverage the images of the criminals in addition to the textual case reports/crime history. Applications built in this way could lead to inadvertent consequences and be used with malicious intent. In this work, we run an audit to investigate the efficiency of standalone VLMs in the bail decision prediction task. We observe that the performance is poor across multiple intersectional groups and models \textit{wrongly deny bail to deserving individuals with very high confidence}. We design different intervention algorithms by first including legal precedents through a RAG pipeline and then fine-tuning the VLMs using innovative schemes. We demonstrate that these interventions substantially improve the performance of bail prediction. Our work paves the way for the design of smarter interventions on VLMs in the future, before they can be deployed for real-world legal judgment prediction.

Detecting anomalies in irregularly sampled multi-variate time-series is challenging, especially in data-scarce settings. Here we introduce an anomaly detection framework for irregularly sampled time-series that leverages neural jump ordinary differential equations (NJODEs). The method infers conditional mean and variance trajectories in a fully path dependent way and computes anomaly scores. On synthetic data containing jump, drift, diffusion, and noise anomalies, the framework accurately identifies diverse deviations. Applied to infant gut microbiome trajectories, it delineates the magnitude and persistence of antibiotic-induced disruptions: revealing prolonged anomalies after second antibiotic courses, extended duration treatments, and exposures during the second year of life. We further demonstrate the predictive capabilities of the inferred anomaly scores in accurately predicting antibiotic events and outperforming diversity-based baselines. Our approach accommodates unevenly spaced longitudinal observations, adjusts for static and dynamic covariates, and provides a foundation for inferring microbial anomalies induced by perturbations, offering a translational opportunity to optimize intervention regimens by minimizing microbial disruptions.

Deep neural networks have been widely adopted in many vision and robotics applications with visual inputs. It is essential to verify its robustness against semantic transformation perturbations, such as brightness and contrast. However, current certified training and robustness certification methods face the challenge of over-parameterization, which hinders the tightness and scalability due to the over-complicated neural networks. To this end, we first analyze stability and variance of layers and neurons against input perturbation, showing that certifiable robustness can be indicated by a fundamental Unbiased and Smooth Neuron metric (USN). Based on USN, we introduce a novel neural network pruning method that removes neurons with low USN and retains those with high USN, thereby preserving model expressiveness without over-parameterization. To further enhance this pruning process, we propose a new Wasserstein distance loss to ensure that pruned neurons are more concentrated across layers. We validate our approach through extensive experiments on the challenging robust keypoint detection task, which involves realistic brightness and contrast perturbations, demonstrating that our method achieves superior robustness certification performance and efficiency compared to baselines.

Foundation models have reshaped AI by unifying fragmented architectures into scalable backbones with multimodal reasoning and contextual adaptation. In parallel, the long-standing notion of AI agents, defined by the sensing-decision-action loop, is entering a new paradigm: with FMs as their cognitive core, agents transcend rule-based behaviors to achieve autonomy, generalization, and self-reflection. This dual shift is reinforced by real-world demands such as autonomous driving, robotics, virtual assistants, and GUI agents, as well as ecosystem advances in embedded hardware, edge computing, mobile deployment platforms, and communication protocols that together enable large-scale deployment. Yet this convergence collides with reality: while applications demand long-term adaptability and real-time interaction, mobile and edge deployments remain constrained by memory, energy, bandwidth, and latency. This creates a fundamental tension between the growing complexity of FMs and the limited resources of deployment environments. This survey provides the first systematic characterization of adaptive, resource-efficient agentic AI systems. We summarize enabling techniques into elastic inference, test-time adaptation, dynamic multimodal integration, and agentic AI applications, and identify open challenges in balancing accuracy-latency-communication trade-offs and sustaining robustness under distribution shifts. We further highlight future opportunities in algorithm-system co-design, cognitive adaptation, and collaborative edge deployment. By mapping FM structures, cognition, and hardware resources, this work establishes a unified perspective toward scalable, adaptive, and resource-efficient agentic AI. We believe this survey can help readers to understand the connections between enabling technologies while promoting further discussions on the fusion of agentic intelligence and intelligent agents.