A Unifying Perspective on Neighbor Embeddings along the Attraction-Repulsion Spectrum

26 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges (2021, background): “Recent studies on DR algorithms shed light on how the loss function affects the rendering of local structure (Böhm et al., 2020), and provide guidance on how to design good loss functions so that the local and global structure can both be preserved simultaneously (Wang et al., 2020b).”
Understanding How Dimension Reduction Tools Work: An Empirical Approach to Deciphering t-SNE, UMAP, TriMAP, and PaCMAP for Data Visualization (2020, methodology): “Böhm et al. (2020) show that UMAP embeddings correspond to t-SNE embeddings with increased attractive forces, and that another algorithm, ForceAtlas2 (Jacomy et al., 2014), has even larger attractive forces.”
Initialization is critical for preserving global data structure in both t-SNE and UMAP (2021, background): “Future research in machine learning is needed to pinpoint the exact mathematical and algorithmic origins of this difference 12 , and future research in single-cell biology is needed to decide which algorithm is more faithful to the underlying biological data.”