SECURE, ROBUST FEDERATED LEARNING SYSTEM BY MULTI-PARTY TYPE HOMOMORPHIC ENCRYPTION AND FEDERATED LEARNING METHOD

• Main Authors: FUJIMOTO TAKUSHI, SYEDERFAN HOSSEINI, OHAZAMA MITSUHARU, ASHISH KHISTI
• Format: Patent
• Language: English; Japanese
• Published: 19.01.2023
• Subjects: ADVERTISING; CALCULATING; CODING OR CIPHERING APPARATUS FOR CRYPTOGRAPHIC OR OTHERPURPOSES INVOLVING THE NEED FOR SECRECY; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; CRYPTOGRAPHY; DISPLAY; EDUCATION; ELECTRIC COMMUNICATION TECHNIQUE; ELECTRICITY; PHYSICS; SEALS; TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION

To improve secrecy and to reduce communication costs without installing a TTP.SOLUTION: In a federated learning system comprises edges and a server, the edges encrypt gradient information by a common public key and transmit the encrypted gradients to the server; the server adds encrypted gradients received from the plurality of edges, to generate an encrypted aggregated gradient and transmits it to the edges; the edges encrypt the encrypted aggregated gradient to generate edge switch shares and transmit them the server; the server adds the edge switch shares received from the plurality of edges, generates encrypted aggregated gradients for decoding, decodes the generated encrypted aggregated gradient for decoding to generate an aggregated gradient, and transmits it to the edges; and the edges performs learning processing of an AI model using the aggregated gradient received from the server.SELECTED DRAWING: Figure 8 【課題】ＴＴＰを設置することなく、秘匿性を向上し、通信コストを低減する。【解決手段】エッジとサーバとを備える連合学習システムであって、エッジは、勾配情報を共通公開鍵で暗号化した暗号化勾配をサーバに送信し、サーバは、複数のエッジから受信した暗号化勾配を加算して、暗号化集約勾配を生成してエッジに送信し、エッジは、暗号化集約勾配を暗号化したエッジスイッチシェアを生成してサーバに送信し、サーバは、複数のエッジから受信したエッジスイッチシェアを加算して、復号用暗号化集約勾配を生成し、生成された復号用暗号化集約勾配を復号化して集約勾配を生成して、エッジに送信し、エッジは、サーバから受信した集約勾配を用いてＡＩモデルを学習する。【選択図】図８