Results 1  10
of
20
A MultiView Embedding Space for Modeling Internet Images, Tags, and their Semantics
 IJCV
"... This paper investigates the problem of modeling Internet images and associated text or tags for tasks such as imagetoimage search, tagtoimage search, and imagetotag search (image annotation). We start with canonical correlation analysis (CCA), a popular and successful approach for mapping vis ..."
Abstract

Cited by 20 (1 self)
 Add to MetaCart
This paper investigates the problem of modeling Internet images and associated text or tags for tasks such as imagetoimage search, tagtoimage search, and imagetotag search (image annotation). We start with canonical correlation analysis (CCA), a popular and successful approach for mapping visual and textual features to the same latent space, and incorporate a third view capturing highlevel image semantics, represented either by a single category or multiple nonmutuallyexclusive concepts. We present two ways to train the threeview embedding: supervised, with the third view coming from groundtruth labels or search keywords; and unsupervised, with semantic themes automatically obtained by clustering the tags. To ensure high accuracy for retrieval tasks while keeping the learning process scalable, we combine multiple strong visual features
Featureaware label space dimension reduction for multilabel classification problem
, 2012
"... Label space dimension reduction (LSDR) is an efficient and effective paradigm for multilabel classification with many classes. Existing approaches to LSDR, such as compressive sensing and principal label space transformation, exploit only the label part of the dataset, but not the feature part. In ..."
Abstract

Cited by 16 (0 self)
 Add to MetaCart
(Show Context)
Label space dimension reduction (LSDR) is an efficient and effective paradigm for multilabel classification with many classes. Existing approaches to LSDR, such as compressive sensing and principal label space transformation, exploit only the label part of the dataset, but not the feature part. In this paper, we propose a novel approach to LSDR that considers both the label and the feature parts. The approach, called conditional principal label space transformation, is based on minimizing an upper bound of the popular Hamming loss. The minimization step of the approach can be carried out efficiently by a simple use of singular value decomposition. In addition, the approach can be extended to a kernelized version that allows the use of sophisticated feature combinations to assist LSDR. The experimental results verify that the proposed approach is more effective than existing ones to LSDR across many realworld datasets. 1
Efficient Multilabel Classification with Many Labels
"... In multilabel classification, each sample can be associated with a set of class labels. When the number of labels grows to the hundreds or even thousands, existing multilabel classification methods often become computationally inefficient. In recent years, a number of remedies have been proposed. ..."
Abstract

Cited by 6 (0 self)
 Add to MetaCart
In multilabel classification, each sample can be associated with a set of class labels. When the number of labels grows to the hundreds or even thousands, existing multilabel classification methods often become computationally inefficient. In recent years, a number of remedies have been proposed. However, they are based either on simple dimension reduction techniques or involve expensive optimization problems. In this paper, we address this problem by selecting a small subset of class labels that can approximately span the original label space. This is performed by an efficient randomized sampling procedure where the sampling probability of each class label reflects its importance among all the labels. Experiments on a number of realworld multilabel data sets with many labels demonstrate the appealing performance and efficiency of the proposed algorithm. 1.
Multilabel subspace ensemble
 in International Conference on Artificial Intelligence and Statistics (AISTATS
, 2012
"... Abstract A challenging problem of multilabel learning is that both the label space and the model complexity will grow rapidly with the increase in the number of labels, and thus makes the available training samples insufficient for training a proper model. In this paper, we eliminate this problem ..."
Abstract

Cited by 4 (2 self)
 Add to MetaCart
(Show Context)
Abstract A challenging problem of multilabel learning is that both the label space and the model complexity will grow rapidly with the increase in the number of labels, and thus makes the available training samples insufficient for training a proper model. In this paper, we eliminate this problem by learning a mapping of each label in the feature space as a robust subspace, and formulating the prediction as finding the group sparse representation of a given instance on the subspace ensemble. We term this approach as "multilabel subspace ensemble (MSE)". In the training stage, the data matrix is decomposed as the sum of several lowrank matrices and a sparse residual via a randomized optimization, where each lowrank part defines a subspace mapped by a label. In the prediction stage, the group sparse representation on the subspace ensemble is estimated by group lasso. Experiments on several benchmark datasets demonstrate the appealing performance of MSE.
Fastxml: a fast, accurate and stable treeclassifier for extreme multilabel learning.
 In KDD,
, 2014
"... ABSTRACT The objective in extreme multilabel classification is to learn a classifier that can automatically tag a data point with the most relevant subset of labels from a large label set. Extreme multilabel classification is an important research problem since not only does it enable the tacklin ..."
Abstract

Cited by 4 (0 self)
 Add to MetaCart
(Show Context)
ABSTRACT The objective in extreme multilabel classification is to learn a classifier that can automatically tag a data point with the most relevant subset of labels from a large label set. Extreme multilabel classification is an important research problem since not only does it enable the tackling of applications with many labels but it also allows the reformulation of ranking problems with certain advantages over existing formulations. Our objective, in this paper, is to develop an extreme multilabel classifier that is faster to train and more accurate at prediction than the stateoftheart Multilabel Random Forest (MLRF) algorithm [2] and the Label Partitioning for Sublinear Ranking (LPSR) algorithm
Hcsearch for multilabel prediction: An empirical study
 In Proceedings of AAAI Conference on Artificial Intelligence (AAAI
, 2014
"... Abstract Multilabel learning concerns learning multiple, overlapping, and correlated classes. In this paper, we adapt a recent structured prediction framework called HCSearch for multilabel prediction problems. One of the main advantages of this framework is that its training is sensitive to the ..."
Abstract

Cited by 4 (2 self)
 Add to MetaCart
(Show Context)
Abstract Multilabel learning concerns learning multiple, overlapping, and correlated classes. In this paper, we adapt a recent structured prediction framework called HCSearch for multilabel prediction problems. One of the main advantages of this framework is that its training is sensitive to the loss function, unlike the other multilabel approaches that either assume a specific loss function or require a manual adaptation to each loss function. We empirically evaluate our instantiation of the HCSearch framework along with many existing multilabel learning algorithms on a variety of benchmarks by employing diverse task loss functions. Our results demonstrate that the performance of existing algorithms tends to be very similar in most cases, and that the HCSearch approach is comparable and often better than all the other algorithms across different loss functions.
Multilabel Classification via Featureaware Implicit Label Space Encoding
"... To tackle a multilabel classification problem with many classes, recently label space dimension reduction (LSDR) is proposed. It encodes the original label space to a lowdimensional latent space and uses a decoding process for recovery. In this paper, we propose a novel method termed FaIE to pe ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
(Show Context)
To tackle a multilabel classification problem with many classes, recently label space dimension reduction (LSDR) is proposed. It encodes the original label space to a lowdimensional latent space and uses a decoding process for recovery. In this paper, we propose a novel method termed FaIE to perform LSDR via Featureaware Implicit label space Encoding. Unlike most previous work, the proposed FaIE makes no assumptions about the encoding process and directly learns a code matrix, i.e. the encoding result of some implicit encoding function, and a linear decoding matrix. To learn both matrices, FaIE jointly maximizes the recoverability of the original label space from the latent space, and the predictability of the latent space from the feature space, thus making itself featureaware. FaIE can also be specified to learn an explicit encoding function, and extended with kernel tricks to handle nonlinear correlations between the feature space and the latent space. Extensive experiments conducted on benchmark datasets well demonstrate its effectiveness. 1.
An Efficient Probabilistic Framework for MultiDimensional Classification
"... The objective of multidimensional classification is to learn a function that accurately maps each data instance to a vector of class labels. Multidimensional classification appears in a wide range of applications including text categorization, gene functionality classification, semantic image la ..."
Abstract

Cited by 3 (3 self)
 Add to MetaCart
(Show Context)
The objective of multidimensional classification is to learn a function that accurately maps each data instance to a vector of class labels. Multidimensional classification appears in a wide range of applications including text categorization, gene functionality classification, semantic image labeling, etc. Usually, in such problems, the class variables are not independent, but rather exhibit conditional dependence relations among them. Hence, the key to the success of multidimensional classification is to effectively model such dependencies and use them to facilitate the learning. In this paper, we propose a new probabilistic approach that represents class conditional dependencies in an effective yet computationally efficient way. Our approach uses a special treestructured Bayesian network model to represent the conditional joint distribution of the class variables given the feature variables. We develop and present efficient algorithms for learning the model from data and for performing exact probabilistic inferences on the model. Extensive experiments on multiple datasets demonstrate that our approach achieves highly competitive results when it is compared to existing stateoftheart methods.
Active Learning for Sparse Bayesian Multilabel Classification
"... We study the problem of active learning for multilabel classification. We focus on the realworld scenario where the average number of positive (relevant) labels per data point is small leading to positive label sparsity. Carrying out mutual information based nearoptimal active learning in this s ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
We study the problem of active learning for multilabel classification. We focus on the realworld scenario where the average number of positive (relevant) labels per data point is small leading to positive label sparsity. Carrying out mutual information based nearoptimal active learning in this setting is a challenging task since the computational complexity involved is exponential in the total number of labels. We propose a novel inference algorithm for the sparse Bayesian multilabel model of [17]. The benefit of this alternate inference scheme is that it enables a natural approximation of the mutual information objective. We prove that the approximation leads to an identical solution to the exact optimization problem but at a fraction of the optimization cost. This allows us to carry out efficient, nonmyopic, and nearoptimal active learning for sparse multilabel classification. Extensive experiments reveal the effectiveness of the method.
Multilabel Classification with Output Kernels
"... Abstract. Although multilabel classification has become an increasingly important problem in machine learning, current approaches remain restricted to learning in the original label space (or in a simple linear projection of the original label space). Instead, we propose to use kernels on output la ..."
Abstract

Cited by 3 (2 self)
 Add to MetaCart
(Show Context)
Abstract. Although multilabel classification has become an increasingly important problem in machine learning, current approaches remain restricted to learning in the original label space (or in a simple linear projection of the original label space). Instead, we propose to use kernels on output label vectors to significantly expand the forms of label dependence that can be captured. The main challenge is to reformulate standard multilabel losses to handle kernels between output vectors. We first demonstrate how a stateoftheart large margin loss for multilabel classification can be reformulated, exactly, to handle output kernels as well as input kernels. Importantly, the preimage problem for multilabel classification can be easily solved at test time, while the training procedure can still be simply expressed as a quadratic program in a dual parameter space. We then develop a projected gradient descent training procedure for this new formulation. Our empirical results demonstrate the efficacy of the proposed approach on complex image labeling tasks. 1