No abstract

Recent secure aggregation protocols enable privacypreserving federated learning for high-dimensional models among thousands or even millions of participants. Due to the scale of these use cases, however, end-to-end empirical evaluation of these protocols is impossible. We present OLYMPIA, a framework for empirical evaluation of secure protocols via simulation. OLYMPIA provides an embedded domain-specific language for defining protocols, and a simulation framework for evaluating their performance. We implement several recent secure aggregation protocols using OLYMPIA, and perform the first empirical comparison of their end-to-end running times. We release OLYMPIA as open source.

The new coronavirus, which emerged in early 2020, caused a major global health crisis in 7 continents. An essential step towards fighting this virus is computed tomography (CT) scans. CT scans are an effective radiological method to detecting the diagnosis in early stage, but have greatly increased the workload of radiologists. For this reason, there are systems needed that will reduce the duration of CT examinations and assist radiologists. In this study, a two-stage system has been proposed for COVID-19 detection. First, a hybrid method is proposed that can segment the infected region from CT images. The reason for this is that there is not always a reference image in the datasets used in the classification. For this purpose; UNet, UNet++, SegNet and PsPNet were used both separately and as hybrids with GAN, to automatically segment infected areas from chest CT slices. According to the segmentation results, cGAN-UNet hybrid system was selected as the most successful method. Experimental results show that the proposed method achieves a segmentation success with a dice score of 92.32% and IoU score of 86.41%. In the second stage, three classifiers which include a Convolutional Neural Network (CNN), a PatchCNN and a Capsule Neural Network (CapsNet) were used to classify the generated masks as either COVID-19 or not, using the segmented images obtained from cGAN-UNet. Success of these classifiers was 99.20%, 92.55% and 73.84%, respectively. According to these results, the highest success was achieved in the system where cGAN-Unet and CNN are used together.

No abstract

As AI-based systems increasingly impact many areas of our lives, auditing these systems for fairness is an increasingly high-stakes problem. Traditional group fairness metrics can miss discrimination against individuals and are difficult to apply after deployment.Counterfactual fairness describes an individualized notion of fairness, but is even more challenging to evaluate after deployment. We present prediction sensitivity, an approach for continual audit of counterfactual fairness in deployed classifiers. Prediction sensitivity helps answer the question: would this prediction have been different, if this individual had belonged to a different demographic group-for every prediction made by the deployed model. Prediction sensitivity can leverage correlations between protected status and other features, and does not require protected status information at prediction time. Our empirical results demonstrate that prediction sensitivity is effective for detecting violations of counterfactual fairness.