AI Paper Atlas

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

Iron: Private Inference on Transformers

Meng Hao, Hongwei Li, Hanxiao Chen, Pengzhi Xing, Guowen Xu, Tianwei Zhang

NeurIPS 2022 · Federated learning and privacy › Privacy attacks and secure learning · approach: Encryption-based private learning · 18 citations (OpenAlex)

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
Compact packing of whole matrix into one ciphertext for HE matmul vs Cheetah row-wise packingIron matmul protocol · Matmul (32,8,16)communication (MB): 2.79 -> 0.11Improved
Customized Softmax/GELU/LayerNorm protocols (merged LayerNorm, square trick) vs SIRNNIron non-linear protocols · Softmaxcommunication (MB): 347.71 -> 206.265Improved
Iron end-to-end vs SIRNN on four BERT modelsIron · BERT end-to-end private inferenceruntime: 3.30x to 11.83x faster; 3.47x to 14.11x less communicationImproved
Iron fixed-point vs plaintext float on BERT-TinyIron · GLUE (MNLI, MRPC, QNLI, SST-2)accuracy: loss not exceeding 0.3%; +0.85% on MNLINo clear effect
Lower fractional scale in fixed-point encoding (below 12) (ablation)Iron · MRPC BERT-Tinyaccuracy: scale 12 matches plaintext; loss under 1% for scale >= 6Hurt

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