TriMap: Large-scale Dimensionality Reduction Using Triplets

By Ehsan Amid, Manfred K. Warmuth. 2019

171 citations recorded by the source corpus.

We introduce ``TriMap''; a dimensionality reduction technique based on triplet constraints that preserves the global accuracy of the data better than the other commonly used methods such as t-SNE, LargeVis, and UMAP. To quantify the global accuracy, we introduce a score which roughly reflects the relative placement of the clusters rather than the individual points. We empirically show the excellent performance of TriMap on a large variety of datasets in terms of the quality of the embedding as well as the runtime. On our performance benchmarks, TriMap easily scales to millions of points without depleting the memory and clearly outperforms t-SNE, LargeVis, and UMAP in terms of runtime.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

How researchers see this paper

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

In Artificial Intelligence

1. Lesia Semenova
2. Chudi Zhong
3. Zhi Chen
4. Haiyang Huang
5. Cynthia Rudin
6. Ser-Nam Lim
7. Yaron Shaposhnik
8. Philipp Berens

In Computer Vision and Pattern Recognition

1. Ser-Nam Lim
2. Cynthia Rudin
3. Zhi Chen
4. Haiyang Huang
5. Lesia Semenova
6. Chudi Zhong
7. Kevin Musgrave
8. Yaron Shaposhnik

In Representation & Self-Supervised Learning

1. J. Böhm
2. Cynthia Rudin
3. Haiyang Huang
4. Dmitry Kobak
5. Philipp Berens
6. Zhi Chen
7. Lesia Semenova
8. Chudi Zhong

Citations in the corpus

• Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges (2021, background): “, 2018), TriMap (Amid and Warmuth, 2019) and PaCMAP (Wang et al.”
• Understanding How Dimension Reduction Tools Work: An Empirical Approach to Deciphering t-SNE, UMAP, TriMAP, and PaCMAP for Data Visualization (2020, methodology): “We used the following implementations for the other DR methods: t-SNE (Belkina et al., 2019), LargeVis (Tang et al., 2016), UMAP (McInnes et al., 2018), and TriMAP (Amid and Warmuth, 2019).”
• A Metric Learning Reality Check (2020, background): “Other applications of embedding losses include learning 3D point cloud features [7], dimensionality reduction for visualization [1], imitation learning [48], sequence prediction [38], and even vanilla image classification [23].”
• A Unifying Perspective on Neighbor Embeddings along the Attraction-Repulsion Spectrum (2020, methodology): “For example, a recently suggested TriMap algorithm (Amid and Warmuth, 2019), which uses negative sampling similar to UMAP, appears to have stronger attractive forces than UMAP (cf. Figure 5 in the original paper), with some TriMap embeddings, e.g. of the Fashion MNIST dataset, looking similar to…”