absorb.md

Daniel

Chronological feed of everything captured from Daniel.

DynaTok: Image-Free 4D Reconstruction from Partial Point Clouds Using Transformers

DynaTok is a novel point-based framework designed for 4D reconstruction from incomplete and unordered partial point cloud sequences, overcoming limitations of existing image-based and object-centric methods. It employs a Transformer-based spatiotemporal encoder to aggregate spatio-temporal information and residual tokens to decouple geometry and motion. A flow-matching decoder then generates complete, temporally coherent 4D point clouds, achieving state-of-the-art results on benchmark datasets.

Generalized Convex Integration Framework for Weak Solutions in Geophysical Fluid Dynamics

This paper introduces a generalized convex integration framework to construct weak solutions for PDEs, particularly those relevant to geophysical fluid dynamics. The method produces $L^\infty$ solutions that are weakly time-continuous. This approach has successfully been applied to yield new results for the incompressible Euler equations, including global existence in $L^\infty$, and is the first convex integration method for hydrostatic Euler equations to construct admissible solutions with natural energy. Additionally, it provides the first existence results for "wild data" solutions in compressible inviscid primitive and inviscid quasi-geostrophic equations.

ECYSAP EYE: Mission-Centric Cyberspace Operations through an SoS Architecture

ECYSAP EYE presents a System-of-Systems (SoS) architecture that moves beyond traditional Cyber Situational Awareness (CySA) by providing mission-relevant artifacts for integration into existing cybersecurity toolchains. It structures a workflow from perception (RCyP) to decision support (WIAR, CySRs, OPRE) and operational execution (DSH, AE, AAR). The design emphasizes explicit integration surfaces to enable incremental deployment and validation, enhancing operational decision-making in mission planning and execution.

MARCIM-WG: Analog-Digital Hybrid Wargame Improves Maritime Cyber Situational Awareness by 34 Percentage Points

The paper proposes MARCIM-WG, a hybrid tabletop wargame that combines physical board elements (tokens, indicators, special cards) with a computational simulation model for adjudication, designed to train decision-makers in maritime cyber defense. Following NATO wargaming methodology, it provides structured decision cycles with measurable consequences, and validation through a comparative study shows a +34.0 percentage-point improvement in the intervention group, with comprehension competencies benefiting the most. The HLD and LLD specifications make the framework replicable for cyber defense training scenarios.