Graph embeddings have gained huge popularity in the recent years as a powerful tool to analyze social networks. However, no prior works have studied potential bias issues inherent within graph embedding. In this paper, we make a first attempt in this direction. In particular, we concentrate on the fairness of node2vec, a popular graph embedding method. Our analyses on two real-world datasets demonstrate the existence of bias in node2vec when used for friendship recommendation. We, therefore, propose a fairness-aware embedding method, namely Fairwalk, which extends node2vec. Experimental results demonstrate that Fairwalk reduces bias under multiple fairness metrics while still preserving the utility.

Objectives:To investigate whether intravascular ultrasound (IVUS) guided percutaneous coronary intervention (PCI) could improve clinical outcomes compared with angiography-guided PCI in the treatment of unprotected left main coronary artery stenosis (ULMCA) in the elderly.Methods:This controlled study was carried out between October 2009 and September 2012, in Qinhuangdao First Hospital, Hebei Province, China. One hundred and twenty-three consecutive patients with ULMCA, aged 70 or older, were randomized to an IVUS-guided group and a control group. The occurrence of major adverse cardiac events (MACE): death, non-fatal myocardial infarction, or target lesion revascularizations) were recorded after 2 years of follow-up.Results:The IVUS-guided group had a lower rate of 2-year MACE than the control group (13.1% versus 29.3%, p=0.031). The incidence of target lesion revascularization was lower in the IVUS-guided group than in the control group (9.1% versus 24%, p=0.045). However, there were no differences in death and myocardial infarction in the 2 groups. On Cox proportional hazard analysis, distal lesion was the independent predictor of MACE (hazard ratio [HR]: 1.99, confidence interval [CI]: 1.129-2.367; p=0.043); IVUS guidance was independent factor of survival free of MACE (HR: 0.414, CI: 0.129-0.867; p=0.033).Conclusion:The use of IVUS could reduce MACE in elderly patients undergoing ULMCA intervention.

In recent years, deep learning has achieved great success in speech enhancement. However, there are two major limitations regarding existing works. First, the Bayesian framework is not adopted in many such deep-learning-based algorithms. In particular, the prior distribution for speech in the Bayesian framework has been shown useful by regularizing the output to be in the speech space, and thus improving the performance. Second, the majority of the existing methods operate on the frequency domain of the noisy speech, such as spectrogram and its variations. The clean speech is then reconstructed using the approach of overlap-add, which is limited by its inherent performance upper bound. This paper presents a Bayesian speech enhancement framework, called BaWN (Bayesian WaveNet), which directly operates on raw audio samples. It adopts the recently announced WaveNet, which is shown to be effective in modeling conditional distributions of speech samples while generating natural speech. Experiments show that BaWN is able to recover clean and natural speech.

No abstract