: Improvements in how KNN is used for massive datasets in AI (like ChatGPT's memory), specifically focusing on Hierarchical Navigable Small World (HNSW) graphs to make searches faster.
However, around that time, there were several significant discussions and "updates" in the machine learning community regarding KNNs, particularly focusing on: KNNs_11-01-23_update.zip
I'd love to dive into the details or help explain any complex parts of it with you! : Improvements in how KNN is used for
: New research on more efficient distance calculations (like Manhattan vs. Euclidean) for high-dimensional data. around that time
: Meta’s Library for efficient similarity search (FAISS) had several optimizations for KNN-related tasks.