{"slug": "incremental-roseland-batch-wise-landmark-diffusion-via-kernelized-sequential-lo", "title": "Incremental Roseland: Batch-wise Landmark Diffusion via Kernelized Sequential Karhunen--Lo\\`eve Update", "summary": "Researchers introduced incremental Roseland (iRoseland), a method that updates the landmark diffusion (Roseland) spectral embedding batch-wise using a kernelized Sequential Karhunen-Loève (SKL) scheme, according to arXiv paper 2611.23414v1. iRoseland requires only a single pass over the data and runs in bounded memory with state independent of the number of points processed, and a forgetting factor lets it track drifting distributions. Experiments show agreement with Roseland to round-off on stationary data and accuracy exceeding equal-memory Roseland by an order of magnitude on long streams.", "body_md": "arXiv:2611.23414v1 Announce Type: new \nAbstract: Landmark diffusion (Roseland) computes a diffusion-map-style spectral embedding from the singular value decomposition of a degree-normalized data-to-landmark kernel matrix, at a cost far below that of full diffusion maps. It assumes, however, that all data are available at once. We introduce incremental Roseland (iRoseland), which casts the landmark affinity as a row-growing matrix whose SVD can be updated with a kernelized Sequential Karhunen--Lo\\`eve (SKL) scheme. Degre correction is handled by a first order expansion about a reference mean kernel row and so that the landmark-side eigenbasis may be recovered from a single landmark-by-landmark eigendecomposition at each step. The presented method requires only a single pass over the data and runs in bounded-memory with state independent of the number of points processed. A forgetting factor lets the method track drifting distributions, and a sample-size-invariant embedding keeps outputs comparable across time. Experiments show agreement with Roseland to round-off on stationary data, and accuracy that exceeds equal-memory Roseland by an order of magnitude on long streams. On drifting data, we observe large gains from forgetting with a window matched to the drift.", "url": "https://wpnews.pro/news/incremental-roseland-batch-wise-landmark-diffusion-via-kernelized-sequential-lo", "canonical_source": "https://arxiv.org/abs/2611.23414", "published_at": "2026-11-19 05:00:00+00:00", "updated_at": "2026-11-19 05:13:02.324068+00:00", "lang": "en", "topics": ["machine-learning", "ai-research", "artificial-intelligence"], "entities": ["Roseland", "iRoseland", "Sequential Karhunen-Loève", "arXiv"], "also_reported_by": [], "takeaway": "iRoseland updates the Roseland landmark-diffusion embedding in a single bounded-memory pass via a kernelized Sequential Karhunen-Loève scheme, matching Roseland to round-off on stationary data and beating equal-memory Roseland by an order of magnitude on long streams.", "recommendation": null, "alternates": {"html": "https://wpnews.pro/news/incremental-roseland-batch-wise-landmark-diffusion-via-kernelized-sequential-lo", "markdown": "https://wpnews.pro/news/incremental-roseland-batch-wise-landmark-diffusion-via-kernelized-sequential-lo.md", "text": "https://wpnews.pro/news/incremental-roseland-batch-wise-landmark-diffusion-via-kernelized-sequential-lo.txt", "jsonld": "https://wpnews.pro/news/incremental-roseland-batch-wise-landmark-diffusion-via-kernelized-sequential-lo.jsonld"}}