xsamgreen

5 ms·

1. 1.▲
   by xsamgreen 6y ago

   Definitely agreed that CPU is a natural baseline. But there are promising results for implementing NTT on GPUs. Here is something recent https://eprint.iacr.org/2021/124 Nothing will beat an ASIC for NTT though.
2. 2.▲
   by xsamgreen 6y ago

   Remote clients are the only ones with the encryption/decryption keys. The compute provider (Intel in this case) only has public keys. If the remote client is using trusted open source software, I wish the compute provider the best of l
3. 3.▲
   by xsamgreen 6y ago

   You may be thinking of Indistinguishability Obfuscation. Current FHE algorithms (e.g. those undergoing standardization: https://homomorphicencryption.org/ ) don't provide IO. The compute provider knows the exact graph t
4. 4.▲
   by xsamgreen 6y ago

   10^5 speed up will get them to 10x slower than plaintext processing of deep neural networks. See page 15 of DARPA's BAA: https://beta.sam.gov/api/prod/opps/v3/opportunities/resource... For simp
5. 5.▲
   by xsamgreen 6y ago

   FHE is undergoing standardization right now. The leading algorithm candidates are based on lattice crypto which is currently quantum-safe: https://homomorphicencryption.org/ Here are some mature and actively-developed HE li