Temporal Dynamic Embedding for Irregularly Sampled Time Series

In several practical applications, particularly healthcare, clinical data of each patient is individually recorded in a database at irregular intervals as required. This causes a sparse and irregularly sampled time series, which makes it difficult to handle as a structured representation of the prerequisites of neural network models. We therefore propose temporal dynamic embedding (TDE), which enables neural network models to receive data that change the number of variables over time. TDE regards each time series variable as an embedding vector evolving over time, instead of a conventional fixed structured representation, which causes a critical missing problem. For each time step, TDE allows for the selective adoption and aggregation of only observed variable subsets and represents the current status of patient based on current observations. The experiment was conducted on three clinical datasets: PhysioNet 2012, MIMIC-III, and PhysioNet 2019. The TDE model performed competitively or better than the imputation-based baseline and several recent state-of-the-art methods with reduced training runtime.

Reliability and Performance Assessment of Federated Learning on Clinical Benchmark Data

As deep learning have been applied in a clinical context, privacy concerns have increased because of the collection and processing of a large amount of personal data. Recently, federated learning (FL) has been suggested to protect personal privacy because it does not centralize data during the training phase. In this study, we assessed the reliability and performance of FL on benchmark datasets including MNIST and MIMIC-III. In addition, we attempted to verify FL on datasets that simulated a realistic clinical data distribution. We implemented FL that uses a client and server architecture and tested client and server FL on modified MNIST and MIMIC-III datasets. FL delivered reliable performance on both imbalanced and extremely skewed distributions (i.e., the difference of the number of patients and the characteristics of patients in each hospital). Therefore, FL can be suitable to protect privacy when applied to medical data.