Mamba Learns in Context: Structure-Aware Domain Generalization for Multi-Task Point Cloud Understanding

Jincen Jiang, Qianyu Zhou, Yuhang Li, Kui Su, Meili Wang, Jian Chang, Jian Jun Zhang, Xuequan Lu, [0.5em] Bournemouth University · 2026

Read Mamba Learns in Context: Structure-Aware Domain Generalization for Multi-Task Point Cloud Understanding by Jincen Jiang, Qianyu Zhou, Yuhang Li, Kui Su, Meili Wang, Jian Chang, Jian Jun Zhang, Xuequan Lu, [0.5em] Bournemouth University (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-2603-20739.txt.

Description

Jiang and co-authors build a Mamba based framework for point cloud tasks across domains. Single task tools often drop in skill when the domain shifts. Transformers cost too much, while plain Mamba orderings drift with viewpoint change. Their design builds sequences from centroid topology and curvature cues. A layered module then steadies reasoning across domains. At test time, a spectral step shifts target features toward source prototypes. They also add MP3DObject, a real scan dataset for tests.

Catalog record

CREATORJincen Jiang, Qianyu Zhou, Yuhang Li, Kui Su, Meili Wang, Jian Chang, Jian Jun Zhang, Xuequan Lu, [0.5em] Bournemouth University
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61595
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.20739

Browse the full library · About the Tower of Babel