Tag
Proposes Big-means++, a simple algorithm that achieves global optimization quality for big data K-means clustering by systematically curating inputs and using sample-induced surrogate landscapes.
This paper presents an unsupervised clustering-based framework using K-Means++ to detect suspicious trading patterns in capital market data, achieving a silhouette score of 0.561 and identifying 2.02% of trades as potentially fraudulent.
PE-means adapts the private evolution algorithm to differentially private k-means clustering, achieving a 20% average improvement in clustering loss over existing methods.
The author shares their work on reducing the cost of multi-vector retrieval by using k-means as top-1 sparse coding. Omar Khattab adds that late-interaction sparse retrieval with neuron-level inverted indexing on unsupervised sparse autoencoders works well.
A single-pass method combines online k-means palette refinement with ordered Bayer dithering, eliminating the separate pixel-mapping step and yielding slight speedups while producing visually interesting results.