This looks super useful.https://twitter.com/fb_engineering/status/1101557803775254528 …
From the paper: "We note that LtHash is *not* collision resistant among multiset inputs." This is because you can add the same input 2^16 times to produce sum = 0 (mod 2^16). There are ways to mitigate this, but their implementation doesn't. Something to be aware of.
-
-
Supporting multisets is referred to in the paper. Here's some related work by
@dfaranha etc al.: https://academic.oup.com/comjnl/article/60/4/476/2608055 … FB didn't need this use case so didn't implement support. Now that it's open source, someone can send a pull request. -
Apologies, didn't mean to misrepresent your work! Hard to fit sufficient nuance in a tweet...
- 1 more reply
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.