As Large Language Model (LLM) agents become increasingly capable of automating complex, multi-step tasks, the need for robust, secure, and predictable architectural patterns is paramount. This paper provides a comprehensive guide to the ``Plan-then-Execute'' (P-t-E) pattern, an agentic design that separates strategic planning from tactical execution. We explore the foundational principles of P-t-E, detailing its core components - the Planner and the Executor - and its architectural advantages in predictability, cost-efficiency, and reasoning quality over reactive patterns like ReAct (Reason + Act). A central focus is placed on the security implications of this design, particularly its inherent resilience to indirect prompt injection attacks by establishing control-flow integrity. We argue that while P-t-E provides a strong foundation, a defense-in-depth strategy is necessary, and we detail essential complementary controls such as the Principle of Least Privilege, task-scoped tool access, and sandboxed code execution. To make these principles actionable, this guide provides detailed implementation blueprints and working code references for three leading agentic frameworks: LangChain (via LangGraph), CrewAI, and AutoGen. Each framework's approach to implementing the P-t-E pattern is analyzed, highlighting unique features like LangGraph's stateful graphs for re-planning, CrewAI's declarative tool scoping for security, and AutoGen's built-in Docker sandboxing. Finally, we discuss advanced patterns, including dynamic re-planning loops, parallel execution with Directed Acyclic Graphs (DAGs), and the critical role of Human-in-the-Loop (HITL) verification, to offer a complete strategic blueprint for architects, developers, and security engineers aiming to build production-grade, resilient, and trustworthy LLM agents.

Chest radiographs (CXRs) are among the most common tests in medicine. Automated image interpretation may reduce radiologists\' workload and expand access to diagnostic expertise. Deep learning multi-task and foundation models have shown strong performance for CXR interpretation but are vulnerable to shortcut learning, where models rely on spurious and off-target correlations rather than clinically relevant features to make decisions. We introduce RoentMod, a counterfactual image editing framework that generates anatomically realistic CXRs with user-specified, synthetic pathology while preserving unrelated anatomical features of the original scan. RoentMod combines an open-source medical image generator (RoentGen) with an image-to-image modification model without requiring retraining. In reader studies with board-certified radiologists and radiology residents, RoentMod-produced images appeared realistic in 93\% of cases, correctly incorporated the specified finding in 89-99\% of cases, and preserved native anatomy comparable to real follow-up CXRs. Using RoentMod, we demonstrate that state-of-the-art multi-task and foundation models frequently exploit off-target pathology as shortcuts, limiting their specificity. Incorporating RoentMod-generated counterfactual images during training mitigated this vulnerability, improving model discrimination across multiple pathologies by 3-19\% AUC in internal validation and by 1-11\% for 5 out of 6 tested pathologies in external testing. These findings establish RoentMod as a broadly applicable tool for probing and correcting shortcut learning in medical AI. By enabling controlled counterfactual interventions, RoentMod enhances the robustness and interpretability of CXR interpretation models and provides a generalizable strategy for improving foundation models in medical imaging.

Specialized machine learning models, regardless of architecture and training, are susceptible to failures in deployment. With their increasing use in high risk situations, the ability to audit these models by determining their operational design domain (ODD) is crucial in ensuring safety and compliance. However, given the high-dimensional input spaces, this process often requires significant human resources and domain expertise. To alleviate this, we introduce \coolname, an LLM-Agent centric framework for automated generation of semantically relevant test cases to search for failure modes in specialized black-box models. By leveraging LLM-Agents as tool orchestrators, we aim to fit a uncertainty-aware failure distribution model on a learned text-embedding manifold by projecting the high-dimension input space to low-dimension text-embedding latent space. The LLM-Agent is tasked with iteratively building the failure landscape by leveraging tools for generating test-cases to probe the model-under-test (MUT) and recording the response. The agent also guides the search using tools to probe uncertainty estimate on the low dimensional manifold. We demonstrate this process in a simple case using models trained with missing digits on the MNIST dataset and in the real world setting of vision-based intruder detection for aerial vehicles.

Significant advancements in AI-driven multimodal medical image diagnosis have led to substantial improvements in ophthalmic disease identification in recent years. However, acquiring paired multimodal ophthalmic images remains prohibitively expensive. While fundus photography is simple and cost-effective, the limited availability of OCT data and inherent modality imbalance hinder further progress. Conventional approaches that rely solely on fundus or textual features often fail to capture fine-grained spatial information, as each imaging modality provides distinct cues about lesion predilection sites. In this study, we propose a novel unpaired multimodal framework \UOPSL that utilizes extensive OCT-derived spatial priors to dynamically identify predilection sites, enhancing fundus image-based disease recognition. Our approach bridges unpaired fundus and OCTs via extended disease text descriptions. Initially, we employ contrastive learning on a large corpus of unpaired OCT and fundus images while simultaneously learning the predilection sites matrix in the OCT latent space. Through extensive optimization, this matrix captures lesion localization patterns within the OCT feature space. During the fine-tuning or inference phase of the downstream classification task based solely on fundus images, where paired OCT data is unavailable, we eliminate OCT input and utilize the predilection sites matrix to assist in fundus image classification learning. Extensive experiments conducted on 9 diverse datasets across 28 critical categories demonstrate that our framework outperforms existing benchmarks.

Tabular data is the foundation of many applications in fields such as finance and healthcare. Although DNNs tailored for tabular data achieve competitive predictive performance, they are blackboxes with little interpretability. We introduce XNNTab, a neural architecture that uses a sparse autoencoder (SAE) to learn a dictionary of monosemantic features within the latent space used for prediction. Using an automated method, we assign human-interpretable semantics to these features. This allows us to represent predictions as linear combinations of semantically meaningful components. Empirical evaluations demonstrate that XNNTab attains performance on par with or exceeding that of state-of-the-art, black-box neural models and classical machine learning approaches while being fully interpretable.

Purpose: We compared the performance of deep learning (DL) and classical machine learning (ML) algorithms for the classification of 24-hour movement behavior into sleep, sedentary, light intensity physical activity (LPA), and moderate-to-vigorous intensity physical activity (MVPA). Methods: Open-access data from 151 adults wearing a wrist-worn accelerometer (Axivity-AX3) was used. Participants were randomly divided into training, validation, and test sets (121, 15, and 15 participants each). Raw acceleration signals were segmented into non-overlapping 10-second windows, and then a total of 104 handcrafted features were extracted. Four DL algorithms-Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (BiLSTM), Gated Recurrent Units (GRU), and One-Dimensional Convolutional Neural Network (1D-CNN)-were trained using raw acceleration signals and with handcrafted features extracted from these signals to predict 24-hour movement behavior categories. The handcrafted features were also used to train classical ML algorithms, namely Random Forest (RF), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), Logistic Regression (LR), Artificial Neural Network (ANN), and Decision Tree (DT) for classifying 24-hour movement behavior intensities. Results: LSTM, BiLSTM, and GRU showed an overall accuracy of approximately 85% when trained with raw acceleration signals, and 1D-CNN an overall accuracy of approximately 80%. When trained on handcrafted features, the overall accuracy for both DL and classical ML algorithms ranged from 70% to 81%. Overall, there was a higher confusion in classification of MVPA and LPA, compared to sleep and sedentary categories. Conclusion: DL methods with raw acceleration signals had only slightly better performance in predicting 24-hour movement behavior intensities, compared to when DL and classical ML were trained with handcrafted features.

Biomedical question answering (QA) poses significant challenges due to the need for precise interpretation of specialized knowledge drawn from a vast, complex, and rapidly evolving corpus. In this work, we explore how large language models (LLMs) can be used for information retrieval (IR), and an ensemble of zero-shot models can accomplish state-of-the-art performance on a domain-specific Yes/No QA task. Evaluating our approach on the BioASQ challenge tasks, we show that ensembles can outperform individual LLMs and in some cases rival or surpass domain-tuned systems - all while preserving generalizability and avoiding the need for costly fine-tuning or labeled data. Our method aggregates outputs from multiple LLM variants, including models from Anthropic and Google, to synthesize more accurate and robust answers. Moreover, our investigation highlights a relationship between context length and performance: while expanded contexts are meant to provide valuable evidence, they simultaneously risk information dilution and model disorientation. These findings emphasize IR as a critical foundation in Retrieval-Augmented Generation (RAG) approaches for biomedical QA systems. Precise, focused retrieval remains essential for ensuring LLMs operate within relevant information boundaries when generating answers from retrieved documents. Our results establish that ensemble-based zero-shot approaches, when paired with effective RAG pipelines, constitute a practical and scalable alternative to domain-tuned systems for biomedical question answering.

The objective of this paper is to significantly reduce the manual workload required from medical professionals in complex 3D segmentation tasks that cannot be yet fully automated. For instance, in radiotherapy planning, organs at risk must be accurately identified in computed tomography (CT) or magnetic resonance imaging (MRI) scans to ensure they are spared from harmful radiation. Similarly, diagnosing age-related degenerative diseases such as sarcopenia, which involve progressive muscle volume loss and strength, is commonly based on muscular mass measurements often obtained from manual segmentation of medical volumes. To alleviate the manual-segmentation burden, this paper introduces an implicit shape prior to segment volumes from sparse slice manual annotations generalized to the multi-organ case, along with a simple framework for automatically selecting the most informative slices to guide and minimize the next interactions. The experimental validation shows the method's effectiveness on two medical use cases: assisted segmentation in the context of at risks organs for brain cancer patients, and acceleration of the creation of a new database with unseen muscle shapes for patients with sarcopenia.

This paper examines the generalization capacity of two state-of-the-art classification and similarity learning models in reliably identifying users based on their motions in various Extended Reality (XR) applications. We developed a novel dataset containing a wide range of motion data from 49 users in five different XR applications: four XR games with distinct tasks and action patterns, and an additional social XR application with no predefined task sets. The dataset is used to evaluate the performance and, in particular, the generalization capacity of the two models across applications. Our results indicate that while the models can accurately identify individuals within the same application, their ability to identify users across different XR applications remains limited. Overall, our results provide insight into current models generalization capabilities and suitability as biometric methods for user verification and identification. The results also serve as a much-needed risk assessment of hazardous and unwanted user identification in XR and Metaverse applications. Our cross-application XR motion dataset and code are made available to the public to encourage similar research on the generalization of motion-based user identification in typical Metaverse application use cases.

The substantial data volumes encountered in modern particle physics and other domains of fundamental physics research allow (and require) the use of increasingly complex data analysis tools and workflows. While the use of machine learning (ML) tools for data analysis has recently proliferated, these tools are typically special-purpose algorithms that rely, for example, on encoded physics knowledge to reach optimal performance. In this work, we investigate a new and orthogonal direction: Using recent progress in large language models (LLMs) to create a team of agents -- instances of LLMs with specific subtasks -- that jointly solve data analysis-based research problems in a way similar to how a human researcher might: by creating code to operate standard tools and libraries (including ML systems) and by building on results of previous iterations. If successful, such agent-based systems could be deployed to automate routine analysis components to counteract the increasing complexity of modern tool chains. To investigate the capabilities of current-generation commercial LLMs, we consider the task of anomaly detection via the publicly available and highly-studied LHC Olympics dataset. Several current models by OpenAI (GPT-4o, o4-mini, GPT-4.1, and GPT-5) are investigated and their stability tested. Overall, we observe the capacity of the agent-based system to solve this data analysis problem. The best agent-created solutions mirror the performance of human state-of-the-art results.