Sparse support vector machines with L0 approximation for ultra-high dimensional omics data

Assessment of Murine Retinal Acuity Ex Vivo Using Multielectrode Array Recordings

Purpose

Visual acuity, measured by resolution of optotypes on a standard eye chart, is a critical clinical test for function of the visual system in humans. Behavioral tests in animals can be used to estimate visual acuity. However, such tests may be limited in the study of mutants or after synthetic vision restoration techniques. Because the total response of the retina to a visual scene is encoded in spiking patterns of retinal ganglion cells, it should be possible to estimate visual acuity in vitro from the retina by analyzing retinal ganglion cell output in response to test stimuli.

Methods

We created a method, EyeCandy, that combines a visual stimulus-generating engine with analysis of multielectrode array retinal recordings via a machine learning approach to measure murine retinal acuity in vitro. Visual stimuli included static checkerboards, drifting gratings, and letter optotypes.

Results

In retinas from wild-type C57Bl/6 mice, retinal acuity measurement for a drifting grating was 0.4 cycles per degree. In contrast, retinas from adult rd1 mice with outer retinal degeneration showed no detectable acuity. A comparison of acuities among different regions of the retina revealed substantial variation, with the inferior-nasal quadrant having highest RA. Letter classification accuracy of a projected Early Treatment Diabetic Retinopathy eye chart reached 99% accuracy for logMAR 3.0 letters. EyeCandy measured a restored RA of 0.05 and 0.08 cycles per degree for static and dynamic stimuli respectively from the retina of the rd1 mouse treated with the azobenzene photoswitch BENAQ.

Conclusions

Machine learning may be used to estimate retinal acuity.

Translational Relevance

The use of ex vivo retinal acuity measurement may allow determination of effects of mutations, drugs, injury, or other manipulations on retinal visual function.

Medical resource allocation planning by integrating machine learning and optimization models

Patients' waiting time is a major issue in the Canadian healthcare system. The planning for resource allocation impacts patients' waiting time in medicare settings. This research focuses on the reduction of patients' waiting time by providing better planning for radiological resource allocation and efficient workload distribution. Resource allocation planning is directly related to the number of patient-arrival and it is hard to predict such uncertain parameters in the future time frame. The number of patient-arrival also varies across different modalities and different timeframes which makes the patient-arrival prediction challenging. In this research, a new three-phase solution framework is proposed where a new multi-target machine learning technique is integrated with an optimization model. In the first phase, a novel Ensemble of Pruned Regressor Chain (EPRC) model is developed and trained offline to predict uncertain parameters, such as patients' arrival. The proposed model is then compared with two popular multi-target prediction methods to evaluate the model's accuracy. In the second phase, the trained model is deployed in the real-time environment to forecast patients' arrival, miss Turn Around Time (miss-TAT) rate, and probable workload count. The forecasted data is used in phase three where a new multi-objective optimization model is developed to determine workload allocation. The Weighted-sum method is used to get efficient solutions. The proposed model is deployed in a Canadian healthcare company and evaluated using real-time healthcare data. It is observed in terms of accuracy, the proposed EPRC model performed 10.81 % better compared to the other multi-target models considered in this study. It is also noticed that the forecasting results have a direct impact on the workload distribution, where the proposed model decreases the total workload by approximately 25 %. Besides, the result shows the efficient workload distribution provided by the proposed framework can reduce the average patients' waiting time by 8.17 %.

Sparse SVM for Sufficient Data Reduction

Kernel-based methods for support vector machines (SVM) have shown highly advantageous performance in various applications. However, they may incur prohibitive computational costs for large-scale sample datasets. Therefore, data reduction (reducing the number of support vectors) appears to be necessary, which gives rise to the topic of the sparse SVM. Motivated by this problem, the sparsity constrained kernel SVM optimization has been considered in this paper in order to control the number of support vectors. Based on the established optimality conditions associated with the stationary equations, a Newton-type method is developed to handle the sparsity constrained optimization. This method is found to enjoy the one-step convergence property if the starting point is chosen to be close to a local region of a stationary point, thereby leading to a super-high computational speed. Numerical comparisons with several powerful solvers demonstrate that the proposed method performs exceptionally well, particularly for large-scale datasets in terms of a much lower number of support vectors and shorter computational time.