{"id":"arxiv/2602.10530","url":"https://arxiv.org/abs/2602.10530","tags":["stat.ML","cs.LG","math.ST"],"score":0,"title":"Generalized Robust Adaptive-Bandwidth Multi-View Manifold Learning in High Dimensions with Noise","domain":"stat.ML","source":"arxiv","authors":["Xiucai Ding","Chao Shen","Hau-Tieng Wu"],"reviews":0,"subject":"stat.ML","summary":"Multiview datasets are common in scientific and engineering applications, yet existing fusion methods offer limited theoretical guarantees, particularly in the presence of heterogeneous and high-dimensional noise. We propose Generalized Robust Adaptive-Bandwidth Multiview Diffusion Maps (GRAB-MDM), a new kernel-based diffusion geometry framework for integrating multiple noisy data sources. The key innovation of GRAB-MDM is a {view}-dependent bandwidth selection strategy that adapts to the geometry and noise level of each view, enabling a stable and principled construction of multiview…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-02-11","reactions":{}}