Deep learning is increasingly being used in high-stake decision making applications that affect individual lives. However, deep learning models might exhibit algorithmic discrimination behaviors with respect to protected groups, potentially posing negative impacts on individuals and society. Therefore, fairness in deep learning has attracted tremendous attention recently. We provide a comprehensive review covering existing techniques to tackle algorithmic fairness problems from the computational perspective. Specifically, we show that interpretability can serve as a useful ingredient, which could be augmented into the biases detection and mitigation pipelines. We also discuss open research problems and future research directions, aiming to push forward the area of fairness in deep learning and build genuinely fair, accountable, and transparent deep learning systems.
Deep learning has shown its great promise in various biomedical image segmentation tasks. Existing models are typically based on U-Net and rely on an encoder-decoder architecture with stacked local operators to aggregate long-range information gradually. However, only using the local operators limits the efficiency and effectiveness. In this work, we propose the non-local U-Nets, which are equipped with flexible global aggregation blocks, for biomedical image segmentation. These blocks can be inserted into U-Net as size-preserving processes, as well as down-sampling and up-sampling layers. We perform thorough experiments on the 3D multimodality isointense infant brain MR image segmentation task to evaluate the non-local U-Nets. Results show that our proposed models achieve top performances with fewer parameters and faster computation.
Presently, visual and quantitative approaches for image-supported diagnosis of dementing disorders rely on regional intensity rather than on connectivity measurements. Here, we test metabolic connectivity for differentiation between Alzheimer's disease and frontotemporal lobar degeneration. Positron emission tomography with 18F-fluorodeoxyglucose was conducted in 47 patients with mild Alzheimer's disease, 52 patients with mild frontotemporal lobar degeneration, and 45 healthy elderly subjects. Sparse inverse covariance estimation and selection were used to identify patterns of metabolic, inter-subject covariance on the basis of 60 regional values. Relative to healthy subjects, significantly more pathological within-lobe connections were found in the parietal lobe of patients with Alzheimer's disease, and in the frontal and temporal lobes of subjects with frontotemporal lobar degeneration. Relative to the frontotemporal lobar degeneration group, more pathological connections between the parietal and temporal lobe were found in the Alzheimer's disease group. The obtained connectivity patterns differentiated between two patients groups with an overall accuracy of 83%. Linear discriminant analysis and univariate methods provided an accuracy of 74% and 69%, respectively. There are characteristic patterns of abnormal metabolic connectivity in mild Alzheimer's disease and frontotemporal lobar degeneration. Such patterns can be utilized for single-subject analyses and might be more accurate in the differential diagnosis of dementing disorders than traditional intensity-based analyses.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.