Relaxed Transfer of Different Classes via Spectral Partition

Shi, Xiaoxiao; Wang, Fan; Yang, Qiang; Ren, Jiangtao

doi:10.1007/978-3-642-04174-7_24

Cited by 9 publications

(8 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As calculating the KL divergence directly from the data can be time consuming, in [25] a more practical solution is proposed, where an approximation is computed based on the output of a clustering algorithm operating on the combined data (source and target data together). More specifically the following definition of Clusteringbased KL divergence is proposed:…”

Section: Clustering-based Transfer Riskmentioning

confidence: 99%

Transferring knowledge across robots: A risk sensitive approach

Ciarfuglia

Valigi

Ricci

2015

Robotics and Autonomous Systems

View full text Add to dashboard Cite

Section: Clustering-based Transfer Riskmentioning

confidence: 99%

Transferring knowledge across robots: A risk sensitive approach

Ciarfuglia

Valigi

Ricci

2015

Robotics and Autonomous Systems

View full text Add to dashboard Cite

“…The theoretical proof of both frameworks are based on strong assumptions on the predictive power of the individual source domains on the target domain data. In [26], a clustering based knowledge transfer was proposed for applications with different class labels across source and target domains, unlike the application addressed in this paper.…”

Section: Related Workmentioning

confidence: 99%

Multi-source domain adaptation and its application to early detection of fatigue

Chattopadhyay

Panchanathan

et al. 2011

Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

Self Cite

View full text Add to dashboard Cite

We consider the characterization of muscle fatigue through noninvasive sensing mechanism such as surface electromyography (SEMG). While changes in the properties of SEMG signals with respect to muscle fatigue have been reported in the literature, the large variation in these signals across different individuals makes the task of modeling and classification of SEMG signals challenging. Indeed, the variation in SEMG parameters from subject to subject creates differences in the data distribution. In this paper, we propose a transfer learning framework based on the multi-source domain adaptation methodology for detecting different stages of fatigue using SEMG signals, that addresses the distribution differences. In the proposed framework, the SEMG data of a subject represent a domain; data from multiple subjects in the training set form the multiple source domains and the test subject data form the target domain. SEMG signals are predominantly different in conditional probability distribution across subjects. The key feature of the proposed framework is a novel weighting scheme that addresses the conditional probability distribution differences across multiple domains (subjects). We have validated the proposed framework on Surface Electromyogram signals collected from 8 people during a fatigue-causing repetitive gripping activity. Comprehensive experiments on the SEMG data set demonstrate that the proposed method improves the classification accuracy by 20% to 30% over the cases without any domain adaptation method and by 13% to 30% over the existing state-of-the-art domain adaptation methods.

show abstract

“…Existing methods for inductive transfer learning may be classified into an instance-based approach [20,25] and a feature-based approach [1,17,22,23]. For instanced-based transfer learning, each example in the source task is re-evaluated before being added to the target task to help the classification process in the target task.…”

Section: Related Workmentioning

confidence: 99%

“…[1] proposed a method to learn a low-dimensional representation underlying multiple tasks. [23] provided a spectralbased solution to transfer the eigenspace that separates the source data to generate a partition of the target data where the process is adjusted by the KL divergence. [17] tried to find a kernel given a set of predefined kernels.…”

Section: Related Workmentioning

confidence: 99%

Feature-based Inductive Transfer Learning through Minimum Encoding

Shao

Suzuki²

2011

Proceedings of the 2011 SIAM International Conference on Data Mining

View full text Add to dashboard Cite

This paper proposes an Extended Minimum Description Length Principle (EMDLP) for feature-based inductive transfer learning, in which both the source and the target data sets contain class labels and relevant features are transferred from the source domain to the target one. Despite numerous works on this topic, few of them have a solid theoretical framework and are parameter-free. Our EMDLP overcomes these flaws and allows us to evaluate the inferiority of the results of transfer learning with the add-sum of the code lengths of five components: the corresponding two hypotheses, the two data sets with the help of the hypotheses, and the set of the transferred features. We design a code book to build the connections between the source and the target tasks. Extensive experiments using both real and artificial data sets show that EMDLP is robust against noise and performs better on the classification accuracy than the state-of-the-art methods.

show abstract

Relaxed Transfer of Different Classes via Spectral Partition

Cited by 9 publications

References 20 publications

Transferring knowledge across robots: A risk sensitive approach

Transferring knowledge across robots: A risk sensitive approach

Multi-source domain adaptation and its application to early detection of fatigue

Feature-based Inductive Transfer Learning through Minimum Encoding

Contact Info

Product

Resources

About