A5075900919- Digital Mental Health Interventions
- Mental Health Research Topics
- Generative Adversarial Networks and Image Synthesis
- Respiratory Support and Mechanisms
- Sepsis Diagnosis and Treatment
- Domain Adaptation and Few-Shot Learning
- Music and Audio Processing
- Context-Aware Activity Recognition Systems
- Diabetes Management and Research
- Smart Grid Energy Management
- Reinforcement Learning in Robotics
- Advanced Bandit Algorithms Research
- Data Stream Mining Techniques
- Mobile Health and mHealth Applications
- Green IT and Sustainability
- Diabetes and associated disorders
- Cardiac Arrest and Resuscitation
- Time Series Analysis and Forecasting
- Machine Learning in Healthcare
- Mobile Crowdsensing and Crowdsourcing
- Evolutionary Algorithms and Applications
- Hemodynamic Monitoring and Therapy
- COVID-19 Clinical Research Studies
- Model Reduction and Neural Networks
- Privacy-Preserving Technologies in Data
Vrije Universiteit Amsterdam
2018-2024
The major application areas of reinforcement learning (RL) have traditionally been game playing and continuous control. In recent years, however, RL has increasingly applied in systems that interact with humans. can personalize digital to make them more relevant individual users. Challenges personalization settings may be different from challenges found traditional RL. An overview work uses for personalization, is lacking. this work, we introduce a framework use it systematic literature...
Abstract Background Reinforcement learning (RL) holds great promise for intensive care medicine given the abundant availability of data and frequent sequential decision-making. But despite emergence promising algorithms, RL driven bedside clinical decision support is still far from reality. Major challenges include trust safety. To help address these issues, we introduce cross off-policy evaluation policy restriction show how detailed analysis may increase interpretability. As an example,...
Information in Electronic Medical Records (EMRs) can be used to generate accurate predictions for the occurrence of a variety health states, which contribute more pro-active interventions. The very nature EMRs does make application off-the-shelf machine learning techniques difficult. In this paper, we study two approaches making that have hardly been compared past: (1) extracting high-level (temporal) features from and building predictive model, (2) defining patient similarity metric...
In recent years, reinforcement learning (RL) has gained traction in the healthcare domain. particular, RL methods have been explored for haemodynamic optimization of septic patients Intensive Care Unit. Most hospitals however, lack data and expertise model development, necessitating transfer models developed using external datasets. This approach assumes generalizability across different patient populations, validity which not previously tested. addition, there is limited knowledge on safety...
For mechanically ventilated critically ill COVID-19 patients, prone positioning has quickly become an important treatment strategy, however, is labor intensive and comes with potential adverse effects. Therefore, identifying which intubated patients will benefit may help allocate resources. From the multi-center Dutch Data Warehouse of ICU from 25 hospitals, we selected all 3619 episodes in 1142 invasively patients. We excluded longer than 24 h. Berlin ARDS criteria were not formally...
Abstract Background The identification of risk factors for adverse outcomes and prolonged intensive care unit (ICU) stay in COVID-19 patients is essential prognostication, determining treatment intensity, resource allocation. Previous studies have determined on admission only, included a limited number predictors. Therefore, using data from the highly granular multicenter Dutch Data Warehouse, we developed machine learning models to identify ICU mortality, ventilator-free days ICU-free...
We introduce an end-to-end reinforcement learning (RL) solution for the problem of sending personalized digital health interventions. Previous work has shown that interventions can be obtained through RL using simple, discrete state information such as recent activity performed. In reality however, features are often not observed, but instead could inferred from noisy, low-level sensor mobile devices (e.g. accelerometers in phones). One first transform raw data into activities, throw away...
Personalization is very powerful in improving the effectiveness of health interventions. Reinforcement learning (RL) algorithms are suitable for these tailored interventions from sequential data collected about individuals. However, can be fragile. The time to learn intervention policies limited as disengagement user occur quickly. Also, e-Health timing crucial before optimal window passes. We present an approach that learns personalization groups users by combining RL and clustering....
Finding well-defined clusters in data represents a fundamental challenge for many data-driven applications, and largely depends on good representation. Drawing literature regarding representation learning, studies suggest that one key characteristic of latent representations is the ability to produce semantically mixed outputs when decoding linear interpolations two representations. We propose Mixing Consistent Deep Clustering method which encourages appear realistic while adding constraint...
While reinforcement learning (RL) has proven to be the approach of choice for tackling many complex problems, it remains challenging develop and deploy RL agents in real-life scenarios successfully. This paper presents pH-RL (personalization e-Health with RL) a general architecture personalization bring health practice. allows various levels applications online batch learning. Furthermore, we provide general-purpose implementation framework that can integrated healthcare applications. We...