Files
openmlsys-zh/chapter_federated_learning/summary.md
Rudy_Tan a6ed0101f7 Added font descriptions in figures and text (#286)
* Create Pic-Instruction

* Delete Pic-Instruction

* Create Requirements and Instructions

* Update Requirements and Instructions

* Rename Requirements and Instructions to Requirements and Instructions.md

* Update Requirements and Instructions.md

* Update Requirements and Instructions.md

* Add files via upload

* Update summary.md

* Update Requirements and Instructions.md

* Modified font in pic template

Modified font in pic template, under the  Pic -> Instruction

* Added font descriptions in figures and text

Added font descriptions in figures and text, modified the PPT template, under the info->Pic_Instruction path.

* Update summary.md

* Update summary.md

Co-authored-by: Dalong <39682259+eedalong@users.noreply.github.com>
2022-04-25 20:28:28 +08:00

727 B

小结

在这一章,我们简单介绍了联邦学习的背景、系统架构、联邦平均算法、隐私加密算法以及实际部署时的挑战。联邦学习是一个新起步的人工智能算法,可以在“数据保护”与“数据孤岛”这两大约束条件下,建立有效的机器学习模型。
此外,由于联邦学习场景的特殊性(端侧数据不上传、安全隐私要求高和数据非独立同分布等特点),使得系统和算法的开发难度更高。如何平衡计算和通讯的开销,如何保证模型不会泄露隐私,以及算法如何在非独立同分布场景下收敛等等,都需要开发人员对实际的联邦学习场景有更深刻的认识。