A5070812258- Machine Learning in Healthcare
- Artificial Intelligence in Healthcare and Education
- Mobile Health and mHealth Applications
- Risk and Portfolio Optimization
- Artificial Intelligence in Healthcare
- Context-Aware Activity Recognition Systems
- Hospital Admissions and Outcomes
- Healthcare Operations and Scheduling Optimization
- Water resources management and optimization
- Physical Activity and Health
- Economic theories and models
- COVID-19 diagnosis using AI
- Stochastic processes and financial applications
- Digital Mental Health Interventions
- Emergency and Acute Care Studies
- Advanced Optimization Algorithms Research
Massachusetts Institute of Technology
2018-2024
Hartford Hospital
2024
Hartford Financial Services (United States)
2024
London Business School
2024
Artificial intelligence (AI) systems hold great promise to improve healthcare over the next decades. Specifically, AI leveraging multiple data sources and input modalities are poised become a viable method deliver more accurate results deployable pipelines across wide range of applications. In this work, we propose evaluate unified Holistic in Medicine (HAIM) framework facilitate generation testing that leverage multimodal inputs. Our approach uses generalizable pre-processing machine...
<h3>Importance</h3> Despite data aggregation and removal of protected health information, there is concern that deidentified physical activity (PA) collected from wearable devices can be reidentified. Organizations collecting or distributing such suggest the aforementioned measures are sufficient to ensure privacy. However, no studies, our knowledge, have been published demonstrate possibility impossibility reidentifying data. <h3>Objective</h3> To evaluate feasibility accelerometer-measured...
Problem definition: Access to accurate predictions of patients' outcomes can enhance medical staff's decision-making, which ultimately benefits all stakeholders in the hospitals. A large hospital network US has been collaborating with academics and consultants predict short-term long-term for inpatients across their seven Methodology/results: We develop machine learning models that probabilities next 24-hr/48-hr discharge intensive care unit transfers, end-of-stay mortality dispositions. All...
Despite the modeling power for problems under uncertainty, robust optimization (RO) and adaptive RO (ARO) can exhibit too conservative solutions in terms of objective value degradation compared with nominal case. One main reasons behind this conservatism is that, many practical applications, uncertain constraints are directly designed as constraint-wise without taking into account couplings over multiple constraints. In paper, we define a coupled uncertainty set intersection between coupling...
Tabular data is essential for applying machine learning tasks across various industries. However, traditional processing methods do not fully utilize all the information available in tables, ignoring important contextual such as column header descriptions. In addition, pre-processing into a tabular format can remain labor-intensive bottleneck model development. This work introduces TabText, and feature extraction framework that extracts from structures. TabText addresses difficulties by...
To promote healthy behaviors, many mobile health applications provide message-based interventions, such as tips, motivational messages, or suggestions for activities. Ideally, the intervention policies should be carefully designed so that users obtain benefits without being overwhelmed by overly frequent messages. As part of HeartSteps physical-activity intervention, receive messages intended to disrupt sedentary behavior. uses an algorithm uniformly spread out daily message budget over...
A collaboration between Hartford Hospital, MIT, and Dynamic Ideas LLC optimizes nurse staffing in emergency department using a two-phase optimization approach. We developed implemented robust models automated scheduling software, reducing costs by 5%–8%, enhancing patient care coverage 8%–25%. This streamlined process is projected to save approximately $720,000 annually, replacing labor-intensive method previously place.
Despite the modeling power for problems under uncertainty, robust optimization (RO) and adaptive (ARO) can exhibit too conservative solutions in terms of objective value degradation compared to nominal case. One main reasons behind this conservatism is that, many practical applications, uncertain constraints are directly designed as constraint-wise without taking into account couplings over multiple constraints. In paper, we define a coupled uncertainty set intersection between coupling set....
Artificial intelligence (AI) systems hold great promise to improve healthcare over the next decades. Specifically, AI leveraging multiple data sources and input modalities are poised become a viable method deliver more accurate results deployable pipelines across wide range of applications. In this work, we propose evaluate unified Holistic in Medicine (HAIM) framework facilitate generation testing that leverage multimodal inputs. Our approach uses generalizable pre-processing machine...