Incremental Real-Time Personalization in Human Activity Recognition Using Domain Adaptive Batch Normalization: Unterschied zwischen den Versionen

Aus IPD-Institutsseminar
Zur Navigation springen Zur Suche springen
(Die Seite wurde neu angelegt: „{{Vortrag |vortragender=Alan Mazankiewicz |email=alan.mazankiewicz@student.kit.edu |vortragstyp=Masterarbeit |betreuer=Klemens Böhm |termin=Institutsseminar/2…“)
 
 
Zeile 4: Zeile 4:
 
|vortragstyp=Masterarbeit
 
|vortragstyp=Masterarbeit
 
|betreuer=Klemens Böhm
 
|betreuer=Klemens Böhm
|termin=Institutsseminar/2020-06-12
+
|termin=Institutsseminar/2020-06-05
 
|kurzfassung=Human Activity Recognition (HAR) from accelerometers is a fundamental problem in ubiquitous computing. Machine learning based recognition models often perform poorly when applied to new users that were not part of the training data. Previous work has addressed this challenge by personalizing general recognition models to the motion pattern of a new user in a static batch setting. The more challenging online setting has received less attention. No samples from the target user are available in advance, but they arrive sequentially. Additionally, the user's motion pattern may change over time. Thus, adapting to new and forgetting old information must be traded off. Finally, the target user should not have to do any work to use the recognition system by labeling activities. Our work addresses this challenges by proposing an unsupervised online domain adaptation algorithm. It works by aligning the feature distribution of all the subjects, sources and target, within deep neural network layers.
 
|kurzfassung=Human Activity Recognition (HAR) from accelerometers is a fundamental problem in ubiquitous computing. Machine learning based recognition models often perform poorly when applied to new users that were not part of the training data. Previous work has addressed this challenge by personalizing general recognition models to the motion pattern of a new user in a static batch setting. The more challenging online setting has received less attention. No samples from the target user are available in advance, but they arrive sequentially. Additionally, the user's motion pattern may change over time. Thus, adapting to new and forgetting old information must be traded off. Finally, the target user should not have to do any work to use the recognition system by labeling activities. Our work addresses this challenges by proposing an unsupervised online domain adaptation algorithm. It works by aligning the feature distribution of all the subjects, sources and target, within deep neural network layers.
 
}}
 
}}

Aktuelle Version vom 5. Juni 2020, 09:47 Uhr

Vortragende(r) Alan Mazankiewicz
Vortragstyp Masterarbeit
Betreuer(in) Klemens Böhm
Termin Fr 5. Juni 2020
Kurzfassung Human Activity Recognition (HAR) from accelerometers is a fundamental problem in ubiquitous computing. Machine learning based recognition models often perform poorly when applied to new users that were not part of the training data. Previous work has addressed this challenge by personalizing general recognition models to the motion pattern of a new user in a static batch setting. The more challenging online setting has received less attention. No samples from the target user are available in advance, but they arrive sequentially. Additionally, the user's motion pattern may change over time. Thus, adapting to new and forgetting old information must be traded off. Finally, the target user should not have to do any work to use the recognition system by labeling activities. Our work addresses this challenges by proposing an unsupervised online domain adaptation algorithm. It works by aligning the feature distribution of all the subjects, sources and target, within deep neural network layers.