Activity knowledge transfer in smart environments

Rashidi, Parisa; Cook, Diane J.

doi:10.1016/j.pmcj.2011.02.007

Cited by 50 publications

(31 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Essentially this eliminates the sensor dimension and only considers the time of day and day of week of the activity. The Manual technique is the mapping technique currently used by most researchers in activity recognition [Cook et al 2012; Rashidi and Cook 2011; van Kasteren et al 2008]. It does not require any labeled data in the target domain, but it does require the manual definition of sensor locations.…”

Section: Resultsmentioning

confidence: 99%

“…Sensors in the source domain are then mapped to similar sensors in the target domain. Rashidi and Cook also map sensors based on location/function but apply additional transfer learning techniques to better align the source and target datasets [Rashidi and Cook 2010; 2011]. Our approach eliminates the need to manually map the feature spaces as this is handled by the algorithm.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Transfer Learning across Feature-Rich Heterogeneous Feature Spaces via Feature-Space Remapping (FSR)

Feuz

Cook

2015

ACM Trans. Intell. Syst. Technol.

Self Cite

View full text Add to dashboard Cite

Transfer learning aims to improve performance on a target task by utilizing previous knowledge learned from source tasks. In this paper we introduce a novel heterogeneous transfer learning technique, Feature- Space Remapping (FSR), which transfers knowledge between domains with different feature spaces. This is accomplished without requiring typical feature-feature, feature instance, or instance-instance co-occurrence data. Instead we relate features in different feature-spaces through the construction of meta-features. We show how these techniques can utilize multiple source datasets to construct an ensemble learner which further improves performance. We apply FSR to an activity recognition problem and a document classification problem. The ensemble technique is able to outperform all other baselines and even performs better than a classifier trained using a large amount of labeled data in the target domain. These problems are especially difficult because in addition to having different feature-spaces, the marginal probability distributions and the class labels are also different. This work extends the state of the art in transfer learning by considering large transfer across dramatically different spaces.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Transfer Learning across Feature-Rich Heterogeneous Feature Spaces via Feature-Space Remapping (FSR)

Feuz

Cook

2015

ACM Trans. Intell. Syst. Technol.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Current works on environmental sensor mainly focus activity recognition on single resident is to improve the accuracy activity model on single resident in smart home environment [30], [33]- [37]. Hence, the mention work using the CASAS datasets as their ground of work [44].…”

Section: Multi Resident Current Methodologies and Approachmentioning

confidence: 99%

Multi Resident Complex Activity Recognition in Smart Home: A Literature Review

Mohamed¹,

Perumal²,

Sulaiman³

et al. 2017

IJSH

View full text Add to dashboard Cite

show abstract

“…In [38], it is proposed a method of transferring the knowledge of learned activities in multiple physical spaces, e.g. homes A and B, to a new target space, e.g.…”

Section: The Aspect Of Intelligencementioning

confidence: 99%

Identifying aspects of Ambient Intelligence through a review of recent developments

Charalampidou

Mouroutsos

Pavlidis

2012

Journal of Advanced Computer Science &Amp; Technology

View full text Add to dashboard Cite

It has already been realized by the scientific and technical community that a new form of technology is going to lead the future technological developments. This technology will be more humancentric and will be more and more "hidden" within everyday-life objects. It will be smarter, personalized, pervasive and ubiquitous. This technology includes what is called Ambient Intelligence (AmI). In this paper, we identify the main aspects of AmI through a review of the recent developments that have been achieved in these aspects of AmI and Ambient Intelligence Environments (AmIEs), as well as point out the problems yet to be solved and the visions of the future.

show abstract

Activity knowledge transfer in smart environments

Cited by 50 publications

References 32 publications

Transfer Learning across Feature-Rich Heterogeneous Feature Spaces via Feature-Space Remapping (FSR)

Transfer Learning across Feature-Rich Heterogeneous Feature Spaces via Feature-Space Remapping (FSR)

Multi Resident Complex Activity Recognition in Smart Home: A Literature Review

Identifying aspects of Ambient Intelligence through a review of recent developments

Contact Info

Product

Resources

About