Autonomous web agents need to operate under externally imposed or human-specified policies while generating long-horizon trajectories. However, little work has examined whether these trajectories comply with such policies, or whether policy violations persist across different contexts such as domains (e.g., shopping or coding websites) and subdomains (e.g., product search and order management in shopping). To address this gap, we introduce PolicyGuardBench, a benchmark of about 60k examples for detecting policy violations in agent trajectories. From diverse agent runs, we generate a broad set of policies and create both within subdomain and cross subdomain pairings with violation labels. In addition to full-trajectory evaluation, PolicyGuardBench also includes a prefix-based violation detection task where models must anticipate policy violations from truncated trajectory prefixes rather than complete sequences. Using this dataset, we train PolicyGuard-4B, a lightweight guardrail model that delivers strong detection accuracy across all tasks while keeping inference efficient. Notably, PolicyGuard-4B generalizes across domains and preserves high accuracy on unseen settings. Together, PolicyGuardBench and PolicyGuard-4B provide the first comprehensive framework for studying policy compliance in web agent trajectories, and show that accurate and generalizable guardrails are feasible at small scales.

Deep learning for medical imaging is hampered by task-specific models that lack generalizability and prognostic capabilities, while existing'universal' approaches suffer from simplistic conditioning and poor medical semantic understanding. T o address these limitations, we introduce DuPLUS, a deep learning framework for efficient multi-modal medical image analysis. DuPLUS introduces a novel vision-language framework that leverages hierarchical semantic prompts for fine-grained control over the analysis task, a capability absent in prior universal models. T o enable extensibility to other medical tasks, it includes a hierarchical, text-controlled architecture driven by a unique dual-prompt mechanism. F or segmentation, DuPLUS is able to generalize across three imaging modalities, ten different anatomically various medical datasets, encompassing more than 30 organs and tumor types. It outperforms the state-of-the-art task-specific and universal models on 8 out of 10 datasets. W e demonstrate extensibility of its text-controlled architecture by seamless integration of electronic health record (EHR) data for prognosis prediction, and on a head and neck cancer dataset, DuPLUS achieved a Concordance Index (CI) of 0.69. Parameter-efficient fine-tuning enables rapid adaptation to new tasks and modalities from varying centers, establishing DuPLUS as a versatile and clinically relevant solution for medical image analysis. The code for this work is made available at: https: //anonymous.4open.science/r/DuPLUS-6C52

In this paper, we design a signalling game-based emergent communication environment to generate state-of-the-art emergent languages in terms of similarity to human language. This is done with hyperparameter optimization, using XferBench as the objective function. XferBench quantifies the statistical similarity of emergent language to human language by measuring its suitability for deep transfer learning to human language. Additionally, we demonstrate the predictive power of entropy on the transfer learning performance of emergent language as well as corroborate previous results on the entropy-minimization properties of emergent communication systems. Finally, we report generalizations regarding what hyperparameters produce more realistic emergent languages, that is, ones which transfer better to human language.

Rapid robot motion generation is critical in Human-Robot Collaboration (HRC) systems, as robots need to respond to dynamic environments in real time by continuously observing their surroundings and replanning their motions to ensure both safe interactions and efficient task execution. Current sampling-based motion planners face challenges in scaling to high-dimensional configuration spaces and often require post-processing to interpolate and smooth the generated paths, resulting in time inefficiency in complex environments. Optimization-based planners, on the other hand, can incorporate multiple constraints and generate smooth trajectories directly, making them potentially more time-efficient. However, optimization-based planners are sensitive to initialization and may get stuck in local minima. In this work, we present a novel learning-based method that utilizes a Flow Matching model conditioned on a single-view point cloud to learn near-optimal solutions for optimization initialization. Our method does not require prior knowledge of the environment, such as obstacle locations and geometries, and can generate feasible trajectories directly from single-view depth camera input. Simulation studies on a UR5e robotic manipulator in cluttered workspaces demonstrate that the proposed generative initializer achieves a high success rate on its own, significantly improves the success rate of trajectory optimization compared with traditional and learning-based benchmark initializers, requires fewer optimization iterations, and exhibits strong generalization to unseen environments.

We present Omni-Embed-Nemotron, a unified multimodal retrieval embedding model developed to handle the increasing complexity of real-world information needs. While Retrieval-Augmented Generation (RAG) has significantly advanced language models by incorporating external knowledge, existing text-based retrievers rely on clean, structured input and struggle with the visually and semantically rich content found in real-world documents such as PDFs, slides, or videos. Recent work such as ColPali has shown that preserving document layout using image-based representations can improve retrieval quality. Building on this, and inspired by the capabilities of recent multimodal models such as Qwen2.5-Omni, we extend retrieval beyond text and images to also support audio and video modalities. Omni-Embed-Nemotron enables both cross-modal (e.g., text - video) and joint-modal (e.g., text - video+audio) retrieval using a single model. We describe the architecture, training setup, and evaluation results of Omni-Embed-Nemotron, and demonstrate its effectiveness in text, image, and video retrieval.

Humans are black boxes -- we cannot observe their neural processes, yet society functions by evaluating verifiable arguments. AI explainability should follow this principle: stakeholders need verifiable reasoning chains, not mechanistic transparency. We propose using structured argumentation to provide a level of explanation and verification neither interpretability nor LLM-generated explanation is able to offer. Our pipeline achieves state-of-the-art 94.44 macro F1 on the AAEC published train/test split (5.7 points above prior work) and $0.81$ macro F1, $\sim$0.07 above previous published results with comparable data setups, for Argumentative MicroTexts relation classification, converting LLM text into argument graphs and enabling verification at each inferential step. We demonstrate this idea on multi-agent risk assessment using the Structured What-If Technique, where specialized agents collaborate transparently to carry out risk assessment otherwise achieved by humans alone. Using Bipolar Assumption-Based Argumentation, we capture support/attack relationships, thereby enabling automatic hallucination detection via fact nodes attacking arguments. We also provide a verification mechanism that enables iterative refinement through test-time feedback without retraining. For easy deployment, we provide a Docker container for the fine-tuned AMT model, and the rest of the code with the Bipolar ABA Python package on GitHub.

Vision-language models (VLMs) have advanced multimodal reasoning but still face challenges in spatial reasoning for 3D scenes and complex object configurations. To address this, we introduce SpatialViLT, an enhanced VLM that integrates spatial features like depth maps, 3D coordinates, and edge maps through a multi-task learning framework. This approach enriches multimodal embeddings with spatial understanding. We propose two variants: SpatialViLT and MaskedSpatialViLT, focusing on full and masked object regions, respectively. Additionally, SpatialEnsemble combines both approaches, achieving state-of-the-art accuracy. Our models excel in spatial reasoning categories such as directional, topological, and proximity relations, as demonstrated on the challenging Visual Spatial Reasoning (VSR) dataset. This work represents a significant step in enhancing the spatial intelligence of AI systems, crucial for advanced multimodal understanding and real-world applications.

Quantitative photoacoustic computed tomography (qPACT) is a promising imaging modality for estimating physiological parameters such as blood oxygen saturation. However, developing robust qPACT reconstruction methods remains challenging due to computational demands, modeling difficulties, and experimental uncertainties. Learning-based methods have been proposed to address these issues but remain largely unvalidated. Virtual imaging (VI) studies are essential for validating such methods early in development, before proceeding to less-controlled phantom or in vivo studies. Effective VI studies must employ ensembles of stochastically generated numerical phantoms that accurately reflect relevant anatomy and physiology. Yet, most prior VI studies for qPACT relied on overly simplified phantoms. In this work, a realistic VI testbed is employed for the first time to assess a representative 3D learning-based qPACT reconstruction method for breast imaging. The method is evaluated across subject variability and physical factors such as measurement noise and acoustic aberrations, offering insights into its strengths and limitations.

A successful deep learning network is highly dependent not only on the training dataset, but the training algorithm used to condition the network for a given task. The loss function, dataset, and tuning of hyperparameters all play an essential role in training a network, yet there is not much discussion on the reliability or reproducibility of a training algorithm. With the rise in popularity of physics-informed loss functions, this raises the question of how reliable one's loss function is in conditioning a network to enforce a particular boundary condition. Reporting the model variation is needed to assess a loss function's ability to consistently train a network to obey a given boundary condition, and provides a fairer comparison among different methods. In this work, a Pix2Pix network predicting the stress fields of high elastic contrast composites is used as a case study. Several different loss functions enforcing stress equilibrium are implemented, with each displaying different levels of variation in convergence, accuracy, and enforcing stress equilibrium across many training sessions. Suggested practices in reporting model variation are also shared.

Self-recognition is a crucial metacognitive capability for AI systems, relevant not only for psychological analysis but also for safety, particularly in evaluative scenarios. Motivated by contradictory interpretations of whether models possess self-recognition (Panickssery et al., 2024; Davidson et al., 2024), we introduce a systematic evaluation framework that can be easily applied and updated. Specifically, we measure how well 10 contemporary larger language models (LLMs) can identify their own generated text versus text from other models through two tasks: binary self-recognition and exact model prediction. Different from prior claims, our results reveal a consistent failure in self-recognition. Only 4 out of 10 models predict themselves as generators, and the performance is rarely above random chance. Additionally, models exhibit a strong bias toward predicting GPT and Claude families. We also provide the first evaluation of model awareness of their own and others' existence, as well as the reasoning behind their choices in self-recognition. We find that the model demonstrates some knowledge of its own existence and other models, but their reasoning reveals a hierarchical bias. They appear to assume that GPT, Claude, and occasionally Gemini are the top-tier models, often associating high-quality text with them. We conclude by discussing the implications of our findings on AI safety and future directions to develop appropriate AI self-awareness.