Yue Zhao

Keun Soo Kim, Yue Zhao, Houk Jang, Sang Yoon Lee, J. M. Kim, Kwang S. Kim, Jong-Hyun Ahn, P. Kim, Jae-Young Choi, B. Hong

Jianjin Shi, Yue Zhao, Kun Wang, Xuyan Shi, Yue Wang, Huanwei Huang, Yinghua Zhuang, T. Cai, Fengchao Wang, F. Shao

Yue Zhao, Meng Li, Liangzhen Lai, Naveen Suda, Damon Civin, V. Chandra

Federated learning enables resource-constrained edge compute devices, such as mobile phones and IoT devices, to learn a shared model for prediction, while keeping the training data local. This decentralized approach to train models provides privacy, security, regulatory and economic benefits. In this work, we focus on the statistical challenge of federated learning when local data is non-IID. We first show that the accuracy of federated learning reduces significantly, by up to 55% for neural networks trained for highly skewed non-IID data, where each client device trains only on a single class of data. We further show that this accuracy reduction can be explained by the weight divergence, which can be quantified by the earth mover's distance (EMD) between the distribution over classes on each device and the population distribution. As a solution, we propose a strategy to improve training on non-IID data by creating a small subset of data which is globally shared between all the edge devices. Experiments show that accuracy can be increased by 30% for the CIFAR-10 dataset with only 5% globally shared data.

Jianjin Shi, Yue Zhao, Yupeng Wang, Wenqing Gao, Jingjin Ding, Peng Li, Liyan Hu, F. Shao

Yunhao Wang, Yue Zhao, Audrey E. Bollas, Yuru Wang, K. Au

Rapid advances in nanopore technologies for sequencing single long DNA and RNA molecules have led to substantial improvements in accuracy, read length and throughput. These breakthroughs have required extensive development of experimental and bioinformatics methods to fully exploit nanopore long reads for investigations of genomes, transcriptomes, epigenomes and epitranscriptomes. Nanopore sequencing is being applied in genome assembly, full-length transcript detection and base modification detection and in more specialized areas, such as rapid clinical diagnoses and outbreak surveillance. Many opportunities remain for improving data quality and analytical approaches through the development of new nanopores, base-calling methods and experimental protocols tailored to particular applications. Au and colleagues outline the field of nanopore sequencing.