Nov 16, 2018
Lopata Hall, Room 101
"Feasibility Study of Monitoring Deterioration of Outpatients Using Multi-modal Data Collected by Wearables"
Adviser: Chenyang Lu
Hospital readmission rate is high for heart failure patients. Early detection of deterioration will help doctors prevent readmissions, thus reducing health care cost and providing patients with just-in-time intervention. Wearable devices (e.g., wristbands and smart watches) provide a convenient technology for continuous outpatient monitoring. In the paper, we explore the feasibility of monitoring outpatients using Fitbit Charge HR wristbands and the potential of machine learning models to predicting clinical deterioration (readmissions and death) among outpatients discharged from the hospital. We developed and piloted a data collection system in a clinical study which involved 25 heart failure patients recently discharged from a hospital. The results from the clinical study demonstrated the feasibility of continuously monitoring outpatients using wristbands. We observed high levels of patient compliance in wearing the wristbands regularly and satisfactory yield, latency and reliability of data collection from the wristbands to a cloud-based database. Finally, we explored a set of machine learning models to predict deterioration based on the Fitbit data. Through 5-fold cross validation, K nearest neighbor achieved the highest accuracy of 0.8800 for identifying patients at risk of deterioration using the health data from the beginning of the monitoring. Machine learning models based on multimodal data (step, sleep and heart rate) significantly outperformed the traditional clinical approach based on LACE index. Moreover, our proposed Weighted Samples One Class SVM model with estimated confidence can reach high accuracy (0.9635) for predicting the deterioration happening in the future using data collected by a sliding window, which indicates the potential for allowing timely intervention.
"Supporting Time-Sensitive Workloads on Partial CPUs"
Adviser: Chenyang Lu
There is increasing interest to migrate real-time and time-sensitive applications to virtualized environments. Existing cloud operators often use dedicated physical CPUs for latency-sensitive applications. However, dedicating physical CPUs to real-time applications can lead to underutilized resources and coarse-grained resource allocation, which diminish the economic benefits of virtualization. While recent advances in real-time virtualization have made a stride to support real-time applications on shared platforms, existing approaches suffer from a number of practical limitations in term of resource efficiency and complexity. We have developed a simple yet efficient resource allocation approach for real-time virtualization that addresses those limitations; it can assign real-time applications to a VCPU that receives a fraction of the CPU resource of a physical CPU and allows non-real-time applications to share the same CPU. We designed scheduling and resource provisioning techniques to achieve three salient capabilities with partial VCPUs: (1) a partial VCPU can be used to achieve the same task schedule as that on a full VCPU, which particularly appealing to time-sensitive applications that desire the same latency profile on a virtualized host as on a dedicated one; (2) a minimal requirement of VCPU bandwidth that allows a real-time VCPU to fully utilize a physical CPU when needed; (3) a simple approach to configure the VCPU parameters with desirable degrees of robustness for their systems. We have implemented our approach on Xen 4.10.0. Experimental results show that the task schedule within a partial VCPU can closely approximate the one in a full VCPU for synthetic workloads as well as case studies running time-sensitive cloud workloads on Redis and Spark Streaming services.