跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 17 天前AI 評分46

e-means 如何以十分之一的迭代次數超越 k-means

e-means: outperforming k-means (in one-tenth the epochs)

AI 導讀

e-means 是一種採用指數移動平均最佳化程序與 mini-batch 更新的 k-means 聚類算法,收斂速度快於 Lloyd k-means,且所得損失更低,經常勝過 k-means++。

正文

Published on Nov 6

·

Submitted by

Bessa

on Nov 10

Authors:

,

,

,

,

,

Abstract

Lloyd's k-means algorithm is a long-running standard for clustering, with applications in data analysis, retrieval, quantization and data curation. In this paper, we introduce an exponential moving average optimization procedure for k-means with mini-batch updates, powered by modern deep learning machinery and new insights. We also establish scaling laws on how decay rate and batch size impact final performance. This gives us a formulaic rule to fix them for multiple problem sizes and embedding datasets, without the need for any hyperparameter sweep. The resulting clustering algorithm, coined e-means, is much faster than Lloyd k-means and converges to lower losses, frequently even beating k-means++. On a wide array of embeddings datasets, it beats the loss of Lloyd k-means at 25 epochs in 1.4 to 2.6 epochs, which translates to 10-45x GPU wall-clock speedups compared to state-of-the-art k-means implementations. As an example, e-means partitions DINO10B, a dataset of 10 billion DINO embeddings, into 16 million clusters in only 2.5 epochs, reaching a loss below Lloyd's at 25 epochs.

View arXiv page View PDF Project page GitHub Add to collection

Get this paper in your agent:

hf papers read 2611.09501

Don't have the latest CLI?

curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2611.09501 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2611.09501 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.09501 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co