-
Optimizing Sepsis Care through Heuristics Methods in Process Mining: A Trajectory Analysis
Authors:
Alireza Bakhshi,
Erfan Hassannayebi,
Amir Hossein Sadeghi
Abstract:
Process mining can help acquire insightful knowledge and heighten the system's performance. In this study, we surveyed the trajectories of 1050 sepsis patients in a regional hospital in the Netherlands from the registration to the discharge phase. Based on this real-world case study, the event log comprises events and activities related to the emergency ward, admission to hospital wards, and disch…
▽ More
Process mining can help acquire insightful knowledge and heighten the system's performance. In this study, we surveyed the trajectories of 1050 sepsis patients in a regional hospital in the Netherlands from the registration to the discharge phase. Based on this real-world case study, the event log comprises events and activities related to the emergency ward, admission to hospital wards, and discharge enriched with data from lab experiments and triage checklists. At first, we aim to discover this process through Heuristics Miner (HM) and Inductive Miner (IM) methods. Then, we analyze a systematic process model based on organizational information and knowledge. Besides, we address conformance checking given medical guidelines for these patients and monitor the related flows on the systematic process model. The results show that HM and IM are inadequate in identifying the relevant process. However, using a systematic process model based on expert knowledge and organizational information resulted in an average fitness of 97.8%, a simplicity of 77.7%, and a generalization of 80.2%. The analyses demonstrate that process mining can shed light on the patient flow in the hospital and inspect the day-to-day clinical performance versus medical guidelines. Also, the process models obtained by the HM and IM methods cannot provide a concrete comprehension of the process structure for stakeholders compared to the systematic process model. The implications of our findings include the potential for process mining to improve the quality of healthcare services, optimize resource allocation, and reduce costs. Our study also highlights the importance of considering expert knowledge and organizational information in developing effective process models.
△ Less
Submitted 24 March, 2023;
originally announced March 2023.
-
Update Compression for Deep Neural Networks on the Edge
Authors:
Bo Chen,
Ali Bakhshi,
Gustavo Batista,
Brian Ng,
Tat-Jun Chin
Abstract:
An increasing number of artificial intelligence (AI) applications involve the execution of deep neural networks (DNNs) on edge devices. Many practical reasons motivate the need to update the DNN model on the edge device post-deployment, such as refining the model, concept drift, or outright change in the learning task. In this paper, we consider the scenario where retraining can be done on the ser…
▽ More
An increasing number of artificial intelligence (AI) applications involve the execution of deep neural networks (DNNs) on edge devices. Many practical reasons motivate the need to update the DNN model on the edge device post-deployment, such as refining the model, concept drift, or outright change in the learning task. In this paper, we consider the scenario where retraining can be done on the server side based on a copy of the DNN model, with only the necessary data transmitted to the edge to update the deployed model. However, due to bandwidth constraints, we want to minimise the transmission required to achieve the update. We develop a simple approach based on matrix factorisation to compress the model update -- this differs from compressing the model itself. The key idea is to preserve existing knowledge in the current model and optimise only small additional parameters for the update which can be used to reconstitute the model on the edge. We compared our method to similar techniques used in federated learning; our method usually requires less than half of the update size of existing methods to achieve the same accuracy.
△ Less
Submitted 21 April, 2022; v1 submitted 8 March, 2022;
originally announced March 2022.
-
Stability and Transparency Analysis of a Bilateral Teleoperation in Presence of Data Loss
Authors:
A. Bakhshi,
H. A. Talebi,
A. A. Suratgar,
M. Abdeetedal
Abstract:
This paper presents a novel approach for stability and transparency analysis for bilateral teleoperation in the presence of data loss in communication media. A new model for data loss is proposed based on a set of periodic continuous pulses and its finite series representation. The passivity of the overall system is shown using wave variable approach including the newly defined model for data loss…
▽ More
This paper presents a novel approach for stability and transparency analysis for bilateral teleoperation in the presence of data loss in communication media. A new model for data loss is proposed based on a set of periodic continuous pulses and its finite series representation. The passivity of the overall system is shown using wave variable approach including the newly defined model for data loss. Simulation results are presented to show the effectiveness of the proposed approach.
△ Less
Submitted 9 November, 2017;
originally announced November 2017.
