Search for: All records

Federated Survival Analysis (FSA) is an emerging Federated Learning (FL) paradigm that enables training survival models on decentralized data while preserving privacy. However, existing FSA approaches largely overlook the potential risk of bias in predictions arising from demographic and censoring disparities across clients' datasets, which impacts the fairness and performance of federated survival models, especially for underrepresented groups. To address this gap, we introduce FairFSA, a novel FSA framework that adapts existing fair survival models to the federated setting. FairFSA jointly trains survival models using distributionally robust optimization, penalizing worst-case errors across subpopulations that exceed a specified probability threshold. Partially observed survival outcomes in clients are reconstructed with federated pseudo values (FPV) before model training to address censoring. Furthermore, we design a weight aggregation strategy by enhancing the FedAvg algorithm with a fairness-aware concordance index-based aggregation method to foster equitable performance distribution across clients. To the best of our knowledge, this is the first work to study and integrate fairness into Federated Survival Analysis. Comprehensive experiments on distributed non-IID datasets demonstrate FairFSA's superiority in fairness and accuracy over state-of-the-art FSA methods, establishing it as a robust FSA approach capable of handling censoring while providing equitable and accurate survival predictions for all subjects. 

Deep learning has achieved state-of-the-art video action recognition (VAR) performance by comprehending action-related features from raw video. However, these models often learn to jointly encode auxiliary view (viewpoints and sensor properties) information with primary action features, leading to performance degradation under novel views and security concerns by revealing sensor types and locations. Here, we systematically study these shortcomings of VAR models and develop a novel approach, VIVAR, to learn view-invariant spatiotemporal action features removing view information. In particular, we leverage contrastive learning to separate actions and jointly optimize adversarial loss that aligns view distributions to remove auxiliary view information in the deep embedding space using the unlabeled synchronous multiview (MV) video to learn view-invariant VAR system. We evaluate VIVAR using our in-house large-scale time synchronous MV video dataset containing 10 actions with three angular viewpoints and sensors in diverse environments. VIVAR successfully captures view-invariant action features, improves inter and intra-action clusters’ quality, and outperforms SoTA models consistently with 8% more accuracy. We additionally perform extensive studies with our datasets, model architectures, multiple contrastive learning, and view distribution alignments to provide VIVAR insights. We open-source our code and dataset to facilitate further research in view-invariant systems. 

The melting of ice sheets significantly contributes to sea level rise, highlighting the crucial need to comprehend the structure of ice for climate benefits. The stratigraphy of ice sheets revealed through ice layer radargrams gives us a window into historical depth-age correlations and accumulation rates. Harnessing this knowledge is not only crucial for interpreting both past and present ice dynamics, especially concerning the Greenland ice sheet, but also for making informed decisions to mitigate the impacts of climate change. Ice layer tracing is prevalently conducted using manual or semi-automatic approaches, requiring significant time and expertise. This study aims to address the need for efficient and precise tracing methods in a two-step process. This is achieved by utilizing an unsupervised annotation method (i.e., ARESELP) to train deep learning models, thereby reducing the need for extensive and time-consuming manual annotations. Four prominent deep learning-based segmentation techniques, namely U-Net, U-Net+VGG19, U-Net+Inception, and Attention U-Net, are benchmarked. Additionally, various thresholding methods such as binary, Otsu, and CLAHE have been explored to achieve optimal enhancement for the true label annotation images. Our preliminary experiments indicate that the combination of attention U-Net with specific processing techniques yields the best performance in terms of the binary IoU metric.