Feature-wise scaling and shifting: Improving the generalization capability of neural networks through capturing independent information of features.

From the perspective of input features, information can be divided into independent information and correlation information. Current neural networks mainly concentrate on the capturing of correlation information through connection weight parameters supplemented by bias parameters. This paper introduces feature-wise scaling and shifting (FwSS) into neural networks for capturing independent information of features, and proposes a new neural network FwSSNet. In the network, a pair of scale and shift parameters is added before each input of each network layer, and bias is removed. The parameters are initialized as 1 and 0, respectively, and trained at separate learning rates, to guarantee the fully capturing of independence and correlation information. The learning rates of FwSS parameters depend on input data and the training speed ratios of adjacent FwSS and connection sublayers, meanwhile those of weight parameters remain unchanged as plain networks. Further, FwSS unifies the scaling and shifting operations in batch normalization (BN), and FwSSNet with BN is established through introducing a preprocessing layer. FwSS parameters except those in the last layer of the network can be simply trained at the same learning rate as weight parameters. Experiments show that FwSS is generally helpful in improving the generalization capability of both fully connected neural networks and deep convolutional neural networks, and FWSSNets achieve higher accuracies on UCI repository and CIFAR-10.

Integrating contraceptive services into existing perinatal care: protocol for a community-based cluster randomised controlled trial in Shanghai, China.

INTRODUCTION: Postpartum contraception is essential to preventing unintended pregnancies and short interpregnancy intervals. The first year after childbirth is a critical period with a high risk of unintended pregnancy and induced abortion. However, the postpartum contraceptive services are weak in China's existing maternal and child healthcare system. We propose to evaluate the effects of integrating postpartum contraceptive services into the existing perinatal care system via a cluster randomised controlled trial. METHODS AND ANALYSIS: This cluster randomised controlled trial involves all 13 communities of Minhang District, Shanghai, China. Communities will be randomly allocated, seven in the intervention group and six in the control group. One thousand and three hundred women, 100 women in each community, will be recruited in the study. Women assigned to the intervention group will receive postpartum contraceptive education and counselling during pregnancy, childbirth hospitalisation, postpartum home visits and the 42-day postpartum clinic check-up. Women in the control group will receive routine antenatal and postpartum care. Participants will be recruited in the first trimester during pregnancy and followed up to 1 year postpartum. The primary outcome is the incidence of unintended pregnancy within 1 year after childbirth. ETHICS AND DISSEMINATION: The trial received ethical approval from the Ethics Committee of Shanghai Minhang District Maternal and Child Health Care Hospital (#[2020]KS-02, #[2020]KS-05, #[2020]KS-05-EX). Results will be published in academic journals and disseminated in multiple formats for the health professionals and the public. TRIAL REGISTRATION NUMBER: ChiCTR2000034603.