News for hackers. Stuff that matters.
Re:Meta-learning neural Bloom filters by geocar on Saturday July 20, 2019 @04:43AM (#20484765)
I can imagine how the secondary (classic) bloom filter can be smaller: It only needs the bits that fail the neural bloom filter that are actually in the set. That means once you're done training the neural bloom filter, you can revisit your training set to find the set of filters that miss the primary so you know what goes in the secondary.

There are no comments on this comment yet.