✳flâneur — a map of the web's best reading

What’s The Story With HNSW?. Exploring the path to fast nearest… | by Ryan McDermott | Towards Data Science

towardsdatascience.com · 3,629 words · saved by 1 readers

Hierarchical Navigable Small World (HNSW) has become popular as one of the best performing approaches for approximate nearest neighbour search. HNSW is a little complex though, and descriptions often lack a complete and intuitive explanation. This post takes a journey through the history of the HNSW idea to help explain what “hierarchical navigable small world” actually means and why it’s effective. A common application of machine learning is nearest neighbour search, which means finding the most similar items* to a target — for example, to recommend items that are similar to a user’s preferences, or to search for items that are similar to a user’s search query. The simple method is to calculate the similarity of every item to the target and return the closest ones. However, if there are a large number of items (maybe millions), this will be slow. Instead, we can use a structure called an index to make things much faster. There is a trade-off, however. Unlike the simple method, indexes

What's The Story With HNSW? | Towards Data Science What’s The Story With HNSW? Exploring the path to fast nearest neighbour search with Hierarchical Navigable Small Worlds Ryan McDermott Feb 25, 2024 17 min read Share Image created by DALL·E 2 with the prompt “A bright abstract expressionist painting of a layered network of dots connected by lines.” Hierarchical Navigable Small World (HNSW) has become popular as one of the best performing approaches for approximate nearest neighbour search. HNSW is a little complex though, and descriptions often lack a complete and intu

Explore this link on the map →

saved by

related reading