HorizonStream: Long-Horizon Attention for Streaming 3D Reconstruction

Chong Cheng, Peilin Tao, Nanjie Yao, Guanzhi Ding, Xianda Chen, Yuansen Du, Xiaoyang Guo, Wei Yin, Weiqiang Ren, Qian Zhang, Zhengqing Chen, Hao Wang, [0.5em] HKUST(GZ), [0.3em] Corresponding author · 2026

Read HorizonStream: Long-Horizon Attention for Streaming 3D Reconstruction by Chong Cheng, Peilin Tao, Nanjie Yao, Guanzhi Ding, Xianda Chen, Yuansen Du, Xiaoyang Guo, Wei Yin, Weiqiang Ren, Qian Zhang, Zhengqing Chen, Hao Wang, [0.5em] HKUST(GZ), [0.3em] Corresponding author (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-23889.txt.

Description

Chong Cheng and colleagues rebuild streaming 3D reconstruction around a simple insight. Geometric evidence lives on mixed timescales, but current systems impose one blunt influence pattern. Their HorizonStream transformer factorizes an evidence influence kernel. It learns per channel decay for long range geometry. It uses local attention with space time position codes for short range matching. Readout tokens recover stable scale and pose. Trained on 48 frame clips, it holds steady past ten thousand frames with constant memory and linear time.

Catalog record

CREATORChong Cheng, Peilin Tao, Nanjie Yao, Guanzhi Ding, Xianda Chen, Yuansen Du, Xiaoyang Guo, Wei Yin, Weiqiang Ren, Qian Zhang, Zhengqing Chen, Hao Wang, [0.5em] HKUST(GZ), [0.3em] Corresponding author
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE55695
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.23889

Browse the full library · About the Tower of Babel