Machine learning is a sub-field of artificial intelligence that uses data to train predictive models, following a two-phase flow of computationally expensive offline training and fast, cheap inference…
Latent spaces serve three distinct roles in machine learning — descriptive, generative, and predictive — according to a technical explainer that includes runnable Python examples. The descriptive role…
A new experiment by Dylan Castillo comparing Matryoshka Representation Learning (MRL) and Principal Component Analysis (PCA) for reducing embedding dimensions found that both methods preserve retrieva…
Principal Component Analysis (PCA) can reduce the dimensionality of embeddings, shrinking memory usage from 14GB for 9 million 384-dimension vectors to a fraction of that by collapsing redundant dimen…
Metric-Aware PCA (MAPCA) parameterizes principal component analysis by a positive-definite metric matrix, with a canonical subfamily interpolating between standard PCA and output whitening. The paper …