Graphs studied in graph theory are usually more sparsely connected than those studied in binary relations.
Also I have a not-too-relevant nerd snipe for you: You have a trillion bitstrings of length 1536 bits each, which you must classify them into a million buckets of approximately a million vectors each. Given a query bitstring, you must be able to identify a small number of buckets (let’s say 10 buckets) that contain most of its neighbours. Distance between two bitstrings is their Hamming distance.
Graphs studied in graph theory are usually more sparsely connected than those studied in binary relations.
Also I have a not-too-relevant nerd snipe for you: You have a trillion bitstrings of length 1536 bits each, which you must classify them into a million buckets of approximately a million vectors each. Given a query bitstring, you must be able to identify a small number of buckets (let’s say 10 buckets) that contain most of its neighbours. Distance between two bitstrings is their Hamming distance.