arXiv:2604.18740v1 Announce Type: new Abstract: Purpose: Automated C-arm positioning ensures timely treatment in patients requiring emergent interventions. When a conventional Deep Learning (DL) approach for C-arm control fails, clinicians must revert to manual operation, resulting in additional delays. Consequently, an agentic C-arm control framework based on multimodal large language models (MLLMs) is highly desirable, as it can incorporate clinician feedback and use reasoning to make adjustments toward more accurate positioning. Skeletal landmark localization is essential for C-arm control, […]

Ver mais

Polymer-based long-acting injectables (LAIs) have transformed the treatment of chronic diseases by enabling controlled drug delivery, thus reducing dosing frequency and extending therapeutic duration. Achieving controlled drug release from LAIs requires extensive optimization of the complex underlying physicochemical properties. Machine learning (ML) can accelerate LAI development by modeling the complex relationships between LAI properties and drug release. However, recent ML studies have provided limited information on key properties that modulate drug release, due to the lack of custom […]

Ver mais

Existing approaches for analyzing neural network activations, such as PCA and sparse autoencoders, rely on strong structural assumptions. Generative models offer an alternative: they can uncover structure without such assumptions and act as priors that improve intervention fidelity. We explore this direction by training diffusion models on one billion residual stream activations, creating "meta-models" that learn the distribution of a network’s internal states. We find that diffusion loss decreases smoothly with compute and reliably predicts downstream utility. In […]

Ver mais

Sepsis remains one of the leading causes of mortality in intensive care units, where timely and accurate treatment decisions can significantly impact patient outcomes. In this work, we propose an interpretable decision support framework. Our system integrates four core components: (1) a clustering-based stratification module that categorizes patients into low, intermediate, and high-risk groups upon ICU admission, using clustering with statistical validation; (2) a synthetic data augmentation pipeline leveraging variational autoencoders (VAE) and diffusion models to enrich underrepresented […]

Ver mais

When I first started learning programming, I focused heavily on syntax, tutorials, and small exercises. I could solve problems, but I struggled to build something meaningful. That changed when I shifted my mindset from learning code to building real projects. In this article, I’ll share practical lessons that helped me grow as a software engineer and how you can apply them to accelerate your own journey. 1. Stop Watching Tutorials, Start Building Tutorials are useful, but they create […]

Ver mais

On Friday evening, February 6, 2026, at around 7 p.m. Seoul time, an employee at Bithumb, South Korea’s second-largest cryptocurrency exchange, was processing payouts for a promotional campaign. Winners were supposed to receive 2,000 Korean won (about $1.37). Instead, the system credited 2,000 bitcoin. Within minutes, 620,000 BTC appeared in customer accounts across the platform, nearly 3% of Bitcoin’s entire 21 million supply. https://x.com/conorfkenny/status/2020801266230690267?s=20&embedable=true At market prices, that was $44 billion. For context, Bithumb’s actual Bitcoin reserves at […]

Ver mais

arXiv:2602.08849v1 Announce Type: new Abstract: The accuracy of machine learning interatomic potentials suffers from reference data that contains numerical noise. Often originating from unconverged or inconsistent electronic-structure calculations, this noise is challenging to identify. Existing mitigation strategies such as manual filtering or iterative refinement of outliers, require either substantial expert effort or multiple expensive retraining cycles, making them difficult to scale to large datasets. Here, we introduce an on-the-fly outlier detection scheme that automatically down-weights noisy samples, without […]

Ver mais

arXiv:2602.02610v1 Announce Type: new Abstract: Consent is an ethical cornerstone of clinical research and healthcare in general. Although the ethical principles of consent – providing information, ensuring comprehension, and ensuring voluntariness – are well-defined, the technological infrastructure remains outdated. Clinicians are responsible for obtaining informed consent from research subjects or patients, and for managing it before, during, and after clinical trials or care, which is a burden for them. The voluntary nature of participating in clinical research or […]

Ver mais