News for hackers. Stuff that matters.
Re:Efficient Logistic Regression on Large Encrypted Data by xjia on Wednesday July 11, 2018 @04:31PM (#17510057)
The encryption procedure doesn't seem to be salted.
Re:Efficient Logistic Regression on Large Encrypted Data by olliej on Wednesday July 11, 2018 @09:58PM (#17511654)
Oh? Is that how they do it? I recall getting confused about the security guarantees of some HE paper the was on HN a while ago, and someone finally explained that you only get the security guarantees if you have randomness injected.