Improving Autism Spectrum Disorder Prediction by Fusion of Multiple Measures of Resting-State Functional MRI Data.

Autism spectrum disorder (ASD) is a lifelong neurodevelopmental condition characterized by social communication, language and behavior impairments. Leveraging deep learning to automatically predict ASD has attracted more and more attention in the medical and machine learning communities. However, how to select effective measure signals for deep learning prediction is still a challenging problem. In this paper, we studied two kinds of measure signals, i.e., regional homogeneity (ReHo) and Craddock 200 (CC200), which both represents homogeneous functional activity, in the framework of deep learning, and designed a new mechanism to effectively joint them for deep learning based ASD prediction. Extensive experiments on the ABIDE dataset provide empirical evidence in support of effectiveness of our method. In particular, we obtained 79% in terms of accuracy by effectively fusing these two kinds of signals, much better than any single-measure model (ReHo SM-model: ∼69% and CC200 SM-model: ∼70%). These results suggest that leveraging multi-measure signals together are effective for ASD prediction.

SFace: Sigmoid-Constrained Hypersphere Loss for Robust Face Recognition.

Deep face recognition has achieved great success due to large-scale training databases and rapidly developing loss functions. The existing algorithms devote to realizing an ideal idea: minimizing the intra-class distance and maximizing the inter-class distance. However, they may neglect that there are also low quality training images which should not be optimized in this strict way. Considering the imperfection of training databases, we propose that intra-class and inter-class objectives can be optimized in a moderate way to mitigate overfitting problem, and further propose a novel loss function, named sigmoid-constrained hypersphere loss (SFace). Specifically, SFace imposes intra-class and inter-class constraints on a hypersphere manifold, which are controlled by two sigmoid gradient re-scale functions respectively. The sigmoid curves precisely re-scale the intra-class and inter-class gradients so that training samples can be optimized to some degree. Therefore, SFace can make a better balance between decreasing the intra-class distances for clean examples and preventing overfitting to the label noise, and contributes more robust deep face recognition models. Extensive experiments of models trained on CASIA-WebFace, VGGFace2, and MS-Celeb-1M databases, and evaluated on several face recognition benchmarks, such as LFW, MegaFace and IJB-C databases, have demonstrated the superiority of SFace.