Shaofeng Cai

TL;DR: This paper proposes Pivot, a novel solution for privacy preserving vertical decision tree training and prediction, ensuring that no intermediate information is disclosed other than those the clients have agreed to release (i.e., the final tree model and the prediction output).

TL;DR: It is uncovered that the architectures generated by NAS algorithms share a common connection pattern, which contributes to their fast convergence, and it is empirically and theoretically shown that the fast convergence is the consequence of smooth loss landscape and accurate gradient information conducted by the common connectionpattern.

TL;DR: Pivot as discussed by the authors is a solution for privacy preserving vertical decision tree training and prediction, ensuring that no intermediate information is disclosed other than those the clients have agreed to release (i.e., the final tree model and the prediction output).

TL;DR: The order of the dropout operations are proposed to be adjusted to address the conflict between the conventional dropout and the batch normalization operation after it, and other structurally more suited dropout variants are examined and introduced for more efficient and effective regularization for CNNs.

TL;DR: In this paper, the authors analyze the impact of 5G on both traditional and emerging technologies and project their view on future research challenges and opportunities, and investigate how 5G can help the development of federated learning.