Transfer Learning on High Variety Domains for Activity Recognition

Ying, Josh Jia-Ching; Lin, Bo-Hau; Tseng, Vincent S.; Hsieh, Sun-Yuan

doi:10.1145/2818869.2818890

Cited by 5 publications

(1 citation statement)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Researchers are applying new trends of ML such as transfer learning which is the ability to transfer knowledge from one model to another to train it with a minimum amount of data and to reduce computational complexity and effort [167]. Similarly, the other emerging ML method is active learning with an objective to reduce learning complexity and computational cost.…”

Section: Rq3 -What Machine Learning Techniques Have Been Used For The...mentioning

confidence: 99%

Open Datasets in Human Activity Recognition Research—Issues and Challenges: A Review

Alam,

McChesney,

Nicholl

et al. 2023

IEEE Sensors J.

View full text Add to dashboard Cite

Huge amounts of data is generated with the emergence of new sensors technologies. Human Activity Recognition (HAR) data sets are generated from cameras, such as video or still images, capturing human behaviour through sensors like gyroscope, Bluetooth, sound sensors, and accelerometers. These generated data sources are collected by the researchers and formed into open data sets. However, these data sets often show issues during data set construction, sharing, searching, that could produce further challenges for reuse of the data by others. The main objective of this research is to explore the current issues and challenges faced by researchers in the HAR domain. A detail literature review was conducted to extract information from the published literature. Similarly, a questionnaire survey was sent to selected researchers having expertise in the HAR domain and who work with open data sets. The main issues and challenges were identified and classified into a hierarchical structure. This research will help HAR researchers to be aware of the current issues and challenges in the field of HAR open data sets. It will help to promote important attributes applicable to many open data sets, such as privacy, anonymity, platform maintenance, data sets descriptions, metadata, environmental conditions, resources, and training, while constructing and sharing new data sets.

show abstract

Section: Rq3 -What Machine Learning Techniques Have Been Used For The...mentioning

confidence: 99%

Open Datasets in Human Activity Recognition Research—Issues and Challenges: A Review

Alam,

McChesney,

Nicholl

et al. 2023

IEEE Sensors J.

View full text Add to dashboard Cite

show abstract

Recent trends in machine learning for human activity recognition—A survey

Ramamurthy

Roy

2018

WIREs Data Min & Knowl

225

View full text Add to dashboard Cite

There has been an upsurge recently in investigating machine learning techniques for activity recognition (AR) problems as they have been very effective in extracting and learning knowledge from the activity datasets. The technique ranges from heuristically derived hand-crafted feature-based traditional machine learning algorithms to the recently developed hierarchically self-evolving feature-based deep learning algorithms. AR continues to remain a challenging problem in uncontrolled smart environments despite the amount of work contributed by the researcher in this field. The complex, volatile, and chaotic nature of the activity data presents numerous challenges that influence the performance of the AR systems in the wild. In this article, we present a comprehensive overview of recent machine learning and data mining techniques generally employed for AR and the underpinning problems and challenges associated with the existing systems. We also articulate the recent advances and state-of-the-art techniques in this domain in an attempt to identify the possible directions for future AR research.

show abstract