§00 · APOD
Cygnus and the Solitary Tree
A lone tree stands in a quiet meadow in Guadalajara, Spain, silhouetted against the Cygnus region rising above like flames in the night sky. This deep night skyscape is a composite of exposures that reveals a range of brightness and color human eyes can't quite see on their own. Spanning over a thousand times the angular size of the full moon, Cygnus sets the sky afire with active star formation where clouds of gas and dust collapse under gravity until nuclear fusion ignites and new stars are born. These stars ionize the surrounding hydrogen gas, causing it to glow crimson, while tendrils of...
2026-03-18 · © 2025 Horacio Lander / AstroHoracio
Text:
Keighley Rockcliffe
(NASA
GSFC,
UMBC CSST,
CRESST II) ·
NASA APOD ↗
§06 · arXiv Dispatch
Research Filed Today
Preprints submitted to arXiv on March 18, 2026. Science before peer review.
01
Recent advances in video diffusion transformers have enabled interactive gaming world models that allow users to explore generated environments over extended horizons. However, existing approaches struggle with precise action control and long-horizon 3D consistency. Most prior wo...
Jisu Nam, Yicong Hong, Chun-Hao Paul Huang et al. (+9)
02
Recent advances in video generation have revealed an unexpected phenomenon: diffusion-based video models exhibit non-trivial reasoning capabilities. Prior work attributes this to a Chain-of-Frames (CoF) mechanism, where reasoning is assumed to unfold sequentially across video fra...
Ruisi Wang, Zhongang Cai, Fanyi Pu et al. (+11)
03
We introduce SegviGen, a framework that repurposes native 3D generative models for 3D part segmentation. Existing pipelines either lift strong 2D priors into 3D via distillation or multi-view mask aggregation, often suffering from cross-view inconsistency and blurred boundaries, ...
Lin Li, Haoran Feng, Zehuan Huang et al. (+8)
04
Large language models (LLMs) with chain-of-thought reasoning achieve state-of-the-art performance across complex problem-solving tasks, but their verbose reasoning traces and large context requirements make them impractical for edge deployment. These challenges include high token...
Yelysei Bondarenko, Thomas Hehn, Rob Hesselink et al. (+15)
05
Monocular 3D scene reconstruction has recently seen significant progress. Powered by the modern neural architectures and large-scale data, recent methods achieve high performance in depth estimation from a single image. Meanwhile, reconstructing and decomposing common scenes into...
Junaid Ahmed Ansari, Ran Ding, Fabio Pizzati et al. (+1)
06
Learning in simulation provides a useful foundation for scaling robotic manipulation capabilities. However, this paradigm often suffers from a lack of data-generation-ready digital assets, in both scale and diversity. In this work, we present ManiTwin, an automated and efficient ...
Kaixuan Wang, Tianxing Chen, Jiawei Liu et al. (+13)
07
Video Super-Resolution (VSR) aims to restore high-quality video frames from low-resolution (LR) estimates, yet most existing VSR approaches behave like black boxes at inference time: users cannot reliably correct unexpected artifacts, but instead can only accept whatever the mode...
Jiongze Yu, Xiangbo Gao, Pooja Verlani et al. (+4)
08
Recent advances in Large Language Models (LLMs) have enabled conversational AI agents to engage in extended multi-turn interactions spanning weeks or months. However, existing memory systems struggle to reason over temporally grounded facts and preferences that evolve across mont...
Sahil Sen, Elias Lumer, Anmol Gulati et al. (+1)
Source: arXiv.org · Cornell University