Abstract
Instead, we propose to directly learn intervention strategies for the low-level sensor data end-to-end using deep neural networks
and RL. We test our novel approach in a self-developed simulation environment which models, and generates, realistic sensor data for daily human activities and show the short-and long-term efficacy of sending personalized physical workout interventions using RL policies. We compare several different input representations and show that learning using raw sensor data is nearly as effective and much more flexible.
| Original language | English |
|---|---|
| Title of host publication | WI '19 - IEEE/WIC/ACM International Conference on Web Intelligence - Proceedings |
| Place of Publication | New York, NY |
| Publisher | ACM |
| Pages | 258-264 |
| Number of pages | 7 |
| ISBN (Electronic) | 9781450369343 |
| DOIs | |
| Publication status | Published - 14 Oct 2019 |
Fingerprint
Keywords
- Personalization
- Health Interventions
- eHealth
- mHealth
- Sensor data
- Advantage Actor-Critic
- Reinforcement Learning
- LSTM
- GANs
Cite this
}
End-to-End Personalization of Digital Health Interventions using Raw Sensor Data with Deep Reinforcement Learning : A comparative study in digital health interventions for behavior change. / El Hassouni, Ali; Hoogendoorn, Mark; Eiben, Agoston E.; van Otterlo, Martijn; Muhonen, Vesa.
WI '19 - IEEE/WIC/ACM International Conference on Web Intelligence - Proceedings. New York, NY : ACM, 2019. p. 258-264.Research output: Chapter in Book / Report / Conference proceeding › Conference contribution › Academic › peer-review
TY - GEN
T1 - End-to-End Personalization of Digital Health Interventions using Raw Sensor Data with Deep Reinforcement Learning
T2 - A comparative study in digital health interventions for behavior change
AU - El Hassouni, Ali
AU - Hoogendoorn, Mark
AU - Eiben, Agoston E.
AU - van Otterlo, Martijn
AU - Muhonen, Vesa
PY - 2019/10/14
Y1 - 2019/10/14
N2 - We introduce an end-to-end reinforcement learning (RL) solution for the problem of sending personalized digital health interventions. Previous work has shown that personalized interventions can be obtained through RL using simple, discrete state information such as the recent activity performed. In reality however, such features are often not observed, but instead could be inferred from noisy, low-level sensor information obtained from mobile devices (e.g. accelerometers in mobile phones). One could first transform such raw data into discrete activities, but that could throw away important details and would require training a classifier to infer these discrete activities which would need a labeled training set.Instead, we propose to directly learn intervention strategies for the low-level sensor data end-to-end using deep neural networksand RL. We test our novel approach in a self-developed simulation environment which models, and generates, realistic sensor data for daily human activities and show the short-and long-term efficacy of sending personalized physical workout interventions using RL policies. We compare several different input representations and show that learning using raw sensor data is nearly as effective and much more flexible.
AB - We introduce an end-to-end reinforcement learning (RL) solution for the problem of sending personalized digital health interventions. Previous work has shown that personalized interventions can be obtained through RL using simple, discrete state information such as the recent activity performed. In reality however, such features are often not observed, but instead could be inferred from noisy, low-level sensor information obtained from mobile devices (e.g. accelerometers in mobile phones). One could first transform such raw data into discrete activities, but that could throw away important details and would require training a classifier to infer these discrete activities which would need a labeled training set.Instead, we propose to directly learn intervention strategies for the low-level sensor data end-to-end using deep neural networksand RL. We test our novel approach in a self-developed simulation environment which models, and generates, realistic sensor data for daily human activities and show the short-and long-term efficacy of sending personalized physical workout interventions using RL policies. We compare several different input representations and show that learning using raw sensor data is nearly as effective and much more flexible.
KW - Personalization
KW - Health Interventions
KW - eHealth
KW - mHealth
KW - Sensor data
KW - Advantage Actor-Critic
KW - Reinforcement Learning
KW - LSTM
KW - GANs
U2 - 10.1145/3350546.3352527
DO - 10.1145/3350546.3352527
M3 - Conference contribution
SP - 258
EP - 264
BT - WI '19 - IEEE/WIC/ACM International Conference on Web Intelligence - Proceedings
PB - ACM
CY - New York, NY
ER -