An Efficient Algorithm for Local Distance Metric Learning. Yang, L., Jin, R., Sukthankar, R., & Liu, Y. January, 2006. abstract bibtex Learning application-specific distance metrics from la- beled data is critical for both statistical classification and information retrieval. Most of the earlier work in this area has focused on finding metrics that simultane- ously optimize compactness and separability in a global sense. Specifically, such distance metrics attempt to keep all of the data points in each class close together while ensuring that data points from different classes are separated. However, particularly when classes ex- hibit multimodal data distributions, these goals conflict and thus cannot be simultaneously satisfied. This paper proposes a Local Distance Metric (LDM) that aims to optimize local compactness and local separability. We present an efficient algorithm that employs eigenvector analysis and bound optimization to learn the LDM from training data in a probabilistic framework. We demon- strate that LDM achieves significant improvements in both classification and retrieval accuracy compared to global distance learning and kernel-based KNN.
@book{yang_efficient_2006,
title = {An {Efficient} {Algorithm} for {Local} {Distance} {Metric} {Learning}.},
abstract = {Learning application-specific distance metrics from la- beled data is critical for both statistical classification and information retrieval. Most of the earlier work in this area has focused on finding metrics that simultane- ously optimize compactness and separability in a global sense. Specifically, such distance metrics attempt to keep all of the data points in each class close together while ensuring that data points from different classes are separated. However, particularly when classes ex- hibit multimodal data distributions, these goals conflict and thus cannot be simultaneously satisfied. This paper proposes a Local Distance Metric (LDM) that aims to optimize local compactness and local separability. We present an efficient algorithm that employs eigenvector analysis and bound optimization to learn the LDM from training data in a probabilistic framework. We demon- strate that LDM achieves significant improvements in both classification and retrieval accuracy compared to global distance learning and kernel-based KNN.},
author = {Yang, Liu and Jin, Rong and Sukthankar, Rahul and Liu, Yi},
month = jan,
year = {2006},
}
Downloads: 0
{"_id":"eqKuEJwLTzuAQwAMs","bibbaseid":"yang-jin-sukthankar-liu-anefficientalgorithmforlocaldistancemetriclearning-2006","downloads":0,"creationDate":"2018-01-22T16:01:10.218Z","title":"An Efficient Algorithm for Local Distance Metric Learning.","author_short":["Yang, L.","Jin, R.","Sukthankar, R.","Liu, Y."],"year":2006,"bibtype":"book","biburl":"https://bibbase.org/zotero/mh_lenguyen","bibdata":{"bibtype":"book","type":"book","title":"An Efficient Algorithm for Local Distance Metric Learning.","abstract":"Learning application-specific distance metrics from la- beled data is critical for both statistical classification and information retrieval. Most of the earlier work in this area has focused on finding metrics that simultane- ously optimize compactness and separability in a global sense. Specifically, such distance metrics attempt to keep all of the data points in each class close together while ensuring that data points from different classes are separated. However, particularly when classes ex- hibit multimodal data distributions, these goals conflict and thus cannot be simultaneously satisfied. This paper proposes a Local Distance Metric (LDM) that aims to optimize local compactness and local separability. We present an efficient algorithm that employs eigenvector analysis and bound optimization to learn the LDM from training data in a probabilistic framework. We demon- strate that LDM achieves significant improvements in both classification and retrieval accuracy compared to global distance learning and kernel-based KNN.","author":[{"propositions":[],"lastnames":["Yang"],"firstnames":["Liu"],"suffixes":[]},{"propositions":[],"lastnames":["Jin"],"firstnames":["Rong"],"suffixes":[]},{"propositions":[],"lastnames":["Sukthankar"],"firstnames":["Rahul"],"suffixes":[]},{"propositions":[],"lastnames":["Liu"],"firstnames":["Yi"],"suffixes":[]}],"month":"January","year":"2006","bibtex":"@book{yang_efficient_2006,\n\ttitle = {An {Efficient} {Algorithm} for {Local} {Distance} {Metric} {Learning}.},\n\tabstract = {Learning application-specific distance metrics from la- beled data is critical for both statistical classification and information retrieval. Most of the earlier work in this area has focused on finding metrics that simultane- ously optimize compactness and separability in a global sense. Specifically, such distance metrics attempt to keep all of the data points in each class close together while ensuring that data points from different classes are separated. However, particularly when classes ex- hibit multimodal data distributions, these goals conflict and thus cannot be simultaneously satisfied. This paper proposes a Local Distance Metric (LDM) that aims to optimize local compactness and local separability. We present an efficient algorithm that employs eigenvector analysis and bound optimization to learn the LDM from training data in a probabilistic framework. We demon- strate that LDM achieves significant improvements in both classification and retrieval accuracy compared to global distance learning and kernel-based KNN.},\n\tauthor = {Yang, Liu and Jin, Rong and Sukthankar, Rahul and Liu, Yi},\n\tmonth = jan,\n\tyear = {2006},\n}\n\n","author_short":["Yang, L.","Jin, R.","Sukthankar, R.","Liu, Y."],"key":"yang_efficient_2006","id":"yang_efficient_2006","bibbaseid":"yang-jin-sukthankar-liu-anefficientalgorithmforlocaldistancemetriclearning-2006","role":"author","urls":{},"metadata":{"authorlinks":{}}},"search_terms":["efficient","algorithm","local","distance","metric","learning","yang","jin","sukthankar","liu"],"keywords":[],"authorIDs":[],"dataSources":["9cexBw6hrwgyZphZZ","iwKepCrWBps7ojhDx"]}