AI Paper Atlas

Research map / Federated learning and privacy / Privacy attacks and secure learning

Falcon: Fast Spectral Inference on Encrypted Data

Qian Lou, Wen-jie Lu, Cheng Hong, Lei Jiang

NeurIPS 2020 · Federated learning and privacy › Privacy attacks and secure learning · approach: Encryption-based private learning

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace second FC layer of LoLa with homomorphic spectral FC (block-circulant, k=8)LoLa MNIST network · MNISTlatency (s): 2.2 -> 1.2Improved
same replacementLoLa MNIST network · MNISTrotations: 225 -> 120Improved
same replacementLoLa MNIST network · MNISTaccuracy (%): 98.95 -> 98.95No clear effect
replace second convolution layer of LoLa with homomorphic convolution (k=16)LoLa CIFAR-10 network · CIFAR-10latency (s): 730 -> 107Improved
same replacementLoLa CIFAR-10 network · CIFAR-10rotations: 52975 -> 7674Improved
same replacementLoLa CIFAR-10 network · CIFAR-10accuracy (%): 76.5 -> 76.5No clear effect
use block-circulant spectral FC in first FC layer instead of LoLa'sLoLa first FC layer · MNISTrotations: 151 -> 46Improved
use tighter encryption modulus (340-bit instead of 440-bit) (ablation)LoLa · MNISTlatency (s): 2.2 -> 2.1Improved
use tighter modulus Q (330-bit) for LoLa (ablation)LoLa · CIFAR-10latency (s): 730 -> 730No clear effect
add homomorphic DFT and inverse DFT to the network (ablation)Falcon · MNISTnoise / modulus Q (bits): 340 -> 430Hurt
use block-circulant matrix in FC and convolution (ablation)Falcon · MNISTnoise bits: ~10-bitImproved
sparse weights and compact encoding versus CryptoNetsnGraph-HE / FCryptoNets · MNISTlatency (s): 205 -> 39.1Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy