Results 1  10
of
20
Distance metric learning for large margin nearest neighbor classification
 In NIPS
, 2006
"... We show how to learn a Mahanalobis distance metric for knearest neighbor (kNN) classification by semidefinite programming. The metric is trained with the goal that the knearest neighbors always belong to the same class while examples from different classes are separated by a large margin. On seven ..."
Abstract

Cited by 688 (15 self)
 Add to MetaCart
(Show Context)
We show how to learn a Mahanalobis distance metric for knearest neighbor (kNN) classification by semidefinite programming. The metric is trained with the goal that the knearest neighbors always belong to the same class while examples from different classes are separated by a large margin. On seven data sets of varying size and difficulty, we find that metrics trained in this way lead to significant improvements in kNN classification—for example, achieving a test error rate of 1.3 % on the MNIST handwritten digits. As in support vector machines (SVMs), the learning problem reduces to a convex optimization based on the hinge loss. Unlike learning in SVMs, however, our framework requires no modification or extension for problems in multiway (as opposed to binary) classification. 1
Regularized Distance Metric Learning: Theory and Algorithm
"... In this paper, we examine the generalization error of regularized distance metric learning. We show that with appropriate constraints, the generalization error of regularized distance metric learning could be independent from the dimensionality, making it suitable for handling high dimensional data. ..."
Abstract

Cited by 30 (2 self)
 Add to MetaCart
(Show Context)
In this paper, we examine the generalization error of regularized distance metric learning. We show that with appropriate constraints, the generalization error of regularized distance metric learning could be independent from the dimensionality, making it suitable for handling high dimensional data. In addition, we present an efficient online learning algorithm for regularized distance metric learning. Our empirical studies with data classification and face recognition show that the proposed algorithm is (i) effective for distance metric learning when compared to the stateoftheart methods, and (ii) efficient and robust for high dimensional data. 1
SemiSupervised Clustering with Metric Learning: An Adaptive Kernel Method
"... Most existing representative works in semisupervised clustering do not sufficiently solve the violation problem of pairwise constraints. On the other hand, traditional kernel methods for semisupervised clustering not only face the problem of manually tuning the kernel parameters due to the fact th ..."
Abstract

Cited by 9 (0 self)
 Add to MetaCart
(Show Context)
Most existing representative works in semisupervised clustering do not sufficiently solve the violation problem of pairwise constraints. On the other hand, traditional kernel methods for semisupervised clustering not only face the problem of manually tuning the kernel parameters due to the fact that no sufficient supervision is provided, but also lack a measure that achieves better effectiveness of clustering. In this paper, we propose an adaptive Semisupervised Clustering Kernel Method based on Metric learning (SCKMM) to mitigate the above problems. Specifically, we first construct an objective function from pairwise constraints to automatically estimate the parameter of the Gaussian kernel. Then, we use pairwise constraintbased Kmeans approach to solve the violation issue of constraints and to cluster the data. Furthermore, we introduce metric learning into nonlinear semisupervised clustering to improve separability of the data for clustering. Finally, we perform clustering and metric learning simultaneously. Experimental results on a number of realworld data sets validate the effectiveness of the proposed method.
Learning Forgiving Hash Functions: Algorithms and
 Large Scale Tests,” International Joint Conference on Artificial Intelligence
, 2007
"... The problem of efficiently finding similar items in a large corpus of highdimensional data points arises in many realworld tasks, such as music, image, and video retrieval. Beyond the scaling difficulties that arise with lookups in large data sets, the complexity in these domains is exacerbated by ..."
Abstract

Cited by 5 (0 self)
 Add to MetaCart
The problem of efficiently finding similar items in a large corpus of highdimensional data points arises in many realworld tasks, such as music, image, and video retrieval. Beyond the scaling difficulties that arise with lookups in large data sets, the complexity in these domains is exacerbated by an imprecise definition of similarity. In this paper, we describe a method to learn a similarity function from only weakly labeled positive examples. Once learned, this similarity function is used as the basis of a hash function to severely constrain the number of points considered for each lookup. Tested on a large realworld audio dataset, only a tiny fraction of the points (~0.27%) are ever considered for each
Online multimodal distance learning for scalable multimedia retrieval
 In WSDM
, 2013
"... In many realword scenarios, e.g., multimedia applications, data often originates from multiple heterogeneous sources or are represented by diverse types of representation, which is often referred to as “multimodal data”. The definition of distance between any two objects/items on multimodal data ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
(Show Context)
In many realword scenarios, e.g., multimedia applications, data often originates from multiple heterogeneous sources or are represented by diverse types of representation, which is often referred to as “multimodal data”. The definition of distance between any two objects/items on multimodal data is a key challenge encountered by many realworld applications, including multimedia retrieval. In this paper, we present a novel online learning framework for learning distance functions on multimodal data through the combination of multiple kernels. In order to attack largescale multimedia applications, we propose Online Multimodal Distance Learning (OMDL) algorithms, which are significantly more efficient and scalable than the stateoftheart techniques. We conducted an extensive set of experiments on multimodal image retrieval applications, in which encouraging results validate the efficacy of the proposed technique.
CluChunk: Clustering Large Scale Usergenerated Content Incorporating Chunklet Information
"... The exponential rise of online content in the form of blogs, microblogs, forums, and multimedia sharing sites has raised an urgent demand for efficient and highquality text clustering algorithms for fast navigation and browsing of users based on better document organization. For several kinds of th ..."
Abstract

Cited by 2 (2 self)
 Add to MetaCart
(Show Context)
The exponential rise of online content in the form of blogs, microblogs, forums, and multimedia sharing sites has raised an urgent demand for efficient and highquality text clustering algorithms for fast navigation and browsing of users based on better document organization. For several kinds of these usergenerated content, it is much easier to obtain the input in small sets, where the data in each set belongs to the same class but with unknown class labels. Such data is viewed as weaklylabeled data and the inherent chunklet information is very useful for improving clustering performance. In this paper, we propose a system CluChunk (clustering chunklet data) to cluster unlabeled web data which incorporates chunklet information. We try to transfer the original feature space by a discriminatively learning linear transformation such that simple unsupervised learning techniques (such as KMeans) in the transformed space can achieve good clustering accuracy. Using larger scale data from some web applications (social media and online forums), we demonstrate that the clustering performance can get significantly improved by: 1)incorporating the inherent weaklylabeled information into the clustering framework; 2)enriching the representation of short text with additional features extracted from the chunklet subset. The proposed approach can be applied to other mining tasks with large scale usergenerated content, like product review summarizing and blog content clustering/classification task.
Classification of WeaklyLabeled Data with Partial Equivalence Relations
"... In many vision problems, instead of having fully labeled training data, it is easier to obtain the input in small groups, where the data in each group is constrained to be from the same class but the actual class label is not known. Such constraints give rise to partial equivalence relations. The ab ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
In many vision problems, instead of having fully labeled training data, it is easier to obtain the input in small groups, where the data in each group is constrained to be from the same class but the actual class label is not known. Such constraints give rise to partial equivalence relations. The absence of class labels prevents the use of standard discriminative methods in this scenario. On the other hand, the stateoftheart techniques that use partial equivalence relations, e.g., Relevant Component Analysis, learn projections that are optimal for data representation, but not discrimination. We show that this leads to poor performance in several realworld applications, especially those with highdimensional data. In this paper, we present a novel discriminative technique for the classification of weaklylabeled data which exploits the nullspace of data scatter matrices to achieve good classification accuracy. We demonstrate the superior performance of both linear and nonlinear versions of our approach on face recognition, clustering, and image retrieval tasks. Results are reported on standard datasets as well as realworld images and videos from the web. 1.
Adaptive distances on sets of vectors
 In ICDM’10. pp.579–588
, 2010
"... Abstract—Recently, there has been a growing interest in learning distances directly from training data. While the previous works focused mainly on adapting distance measures over vectorial data, it is a wellknown fact that many realworld data could not be easily represented as fixed length tuples ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Abstract—Recently, there has been a growing interest in learning distances directly from training data. While the previous works focused mainly on adapting distance measures over vectorial data, it is a wellknown fact that many realworld data could not be easily represented as fixed length tuples of constants. In this paper we address this limitation and propose a novel class of distance learning techniques for learning problems in which instances are set of vectors; examples of such problems include, among others, automatic image annotation and graph classification. We investigate the behavior of the adaptive set distances on a number of artificial and realworld problems and demonstrate that they improve over the standard set distances. Keywordsdistance learning; adaptive distances; complex objects; sets; graphs; I.
Efficient MaxMargin Metric Learning
"... Efficient learning of an appropriate distance metric is an increasingly important problem in machine learning. However, current methods are limited by scalability issues or are unsuited to use with general similarity/dissimilarity constraints. In this paper, we propose an efficient metric learning m ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
Efficient learning of an appropriate distance metric is an increasingly important problem in machine learning. However, current methods are limited by scalability issues or are unsuited to use with general similarity/dissimilarity constraints. In this paper, we propose an efficient metric learning method based on the maxmargin framework with pairwise constraints that has a strong generalization guarantee. First, we reformulate the maxmargin metric learning problem as a structured support vector machine which we can optimize in linear time via a cuttingplane method. Second, we propose a kernelized extention and an approximation method based on matching pursuit that allows lineartime training even in the kernel case.We find our method to be comparable to or better than state of the art metric learning techniques at a number of machine learning and computer vision classification tasks.