Review of multimodal machine learning approaches in healthcare

Machine learning methods in healthcare have traditionally focused on using data from a single modality, limiting their ability to effectively replicate the clinical practice of integrating multiple sources of information for improved decision making. Clinicians typically rely on a variety of data sources including patients' demographic information, laboratory data, vital signs and various imaging data modalities to make informed decisions and contextualise their findings. Recent advances in machine learning have facilitated the more efficient incorporation of multimodal data, resulting in applications that better represent the clinician's approach. Here, we provide a review of multimodal machine learning approaches in healthcare, offering a comprehensive overview of recent literature. We discuss the various data modalities used in clinical diagnosis, with a particular emphasis on imaging data. We evaluate fusion techniques, explore existing multimodal datasets and examine common training strategies.

LT-ViT: A Vision Transformer for multi-label Chest X-ray classification

Vision Transformers (ViTs) are widely adopted in medical imaging tasks, and some existing efforts have been directed towards vision-language training for Chest X-rays (CXRs). However, we envision that there still exists a potential for improvement in vision-only training for CXRs using ViTs, by aggregating information from multiple scales, which has been proven beneficial for non-transformer networks. Hence, we have developed LT-ViT, a transformer that utilizes combined attention between image tokens and randomly initialized auxiliary tokens that represent labels. Our experiments demonstrate that LT-ViT (1) surpasses the state-of-the-art performance using pure ViTs on two publicly available CXR datasets, (2) is generalizable to other pre-training methods and therefore is agnostic to model initialization, and (3) enables model interpretability without grad-cam and its variants.

A Sensitivity Matrix Approach Using Two-Stage Optimization for Voltage Regulation of LV Networks with High PV Penetration

The occurrence of voltage violations are a major deterrent for absorbing more roof-top solar power to smart Low Voltage Distribution Grids (LVDG). Recent studies have focused on decentralized control methods to solve this problem due to the high computational time in performing load flows in centralized control techniques. To address this issue a novel sensitivity matrix is developed to estimate voltages of the network by replacing load flow simulations. In this paper, a Centralized Active, Reactive Power Management System (CARPMS) is proposed to optimally utilize the reactive power capability of smart photo-voltaic inverters with minimal active power curtailment to mitigate the voltage violation problem. The developed sensitivity matrix is able to reduce the time consumed by 48% compared to load flow simulations, enabling near real-time control optimization. Given the large solution space of power systems, a novel two-stage optimization is proposed, where the solution space is narrowed down by a Feasible Region Search (FRS) step, followed by Particle Swarm Optimization (PSO). The performance of the proposed methodology is analyzed in comparison to the load flow method to demonstrate the accuracy and the capability of the optimization algorithm to mitigate voltage violations in near real-time. The deviation of mean voltages of the proposed methodology from load flow method was; 6.5*10^-3 p.u for reactive power control using Q-injection, 1.02*10^-2 p.u for reactive power control using Q-absorption, and 0 p.u for active power curtailment case.

A generalized forecasting solution to enable future insights of COVID-19 at sub-national level resolutions

COVID-19 continues to cause a significant impact on public health. To minimize this impact, policy makers undertake containment measures that however, when carried out disproportionately to the actual threat, as a result if errorneous threat assessment, cause undesirable long-term socio-economic complications. In addition, macro-level or national level decision making fails to consider the localized sensitivities in small regions. Hence, the need arises for region-wise threat assessments that provide insights on the behaviour of COVID-19 through time, enabled through accurate forecasts. In this study, a forecasting solution is proposed, to predict daily new cases of COVID-19 in regions small enough where containment measures could be locally implemented, by targeting three main shortcomings that exist in literature; the unreliability of existing data caused by inconsistent testing patterns in smaller regions, weak deploy-ability of forecasting models towards predicting cases in previously unseen regions, and model training biases caused by the imbalanced nature of data in COVID-19 epi-curves. Hence, the contributions of this study are three-fold; an optimized smoothing technique to smoothen less deterministic epi-curves based on epidemiological dynamics of that region, a Long-Short-Term-Memory (LSTM) based forecasting model trained using data from select regions to create a representative and diverse training set that maximizes deploy-ability in regions with lack of historical data, and an adaptive loss function whilst training to mitigate the data imbalances seen in epi-curves. The proposed smoothing technique, the generalized training strategy and the adaptive loss function largely increased the overall accuracy of the forecast, which enables efficient containment measures at a more localized micro-level.