In many applications non-metric distances are better than metric distances in reflecting the perceptual distances of human beings. Previous studies on non-metric distances mainly focused on supervised setting and did not consider the usefulness of unlabeled data. In this paper, we present probably the first study of label propagation on graphs induced from non-metric distances. The challenge here lies in the fact that the triangular inequality does not hold for non-metric distances and therefore, a direct application of existing label propagation methods will lead to inconsistency and conflict. We show that by applying spectrum transformation, non-metric distances can be converted into metric ones, and thus label propagation can be executed. Such methods, however, suffer from the change of original semantic relations. As a main result of this paper, we prove that any non-metric distance matrix can be decomposed into two metric distance matrices containing different information of the data. Based on this recognition, our proposed NMLP method derives two graphs from the original non-metric distance and performs a joint label propagation on the joint graph. Experiments validate the effectiveness of the proposed NMLP method.

PCA can be smarter and makes more sensible projections. In this paper, we propose smart PCA, an extension to standard PCA to regularize and incorporate external knowledge into model estimation. Based on the probabilistic interpretation of PCA, the inverse Wishart distribution can be used as the informative conjugate prior for the population covariance, and useful knowledge is carried by the prior hyperparameters. We design the hyperparameters to smoothly combine the information from both the domain knowledge and the data itself. The Bayesian point estimation of principal components is in closed form. In empirical studies, smart PCA shows clear improvement on three different criteria: image reconstruction errors, the perceptual quality of the reconstructed images, and the pattern recognition performance.

In this paper, we propose a general formulation for kernel nonnegative matrix factorization with flexible kernels. Specifically, we propose the Gaussian nonnegative matrix factorization (GNMF) algorithm by using the Gaussian kernel in the framework. Different from a recently developed polynomial NMF (PNMF), GNMF finds basis vectors in the kernel-induced feature space and the computational cost is independent of input dimensions. Furthermore, we prove the convergence and nonnegativity of decomposition of our method. Extensive experiments compared with PNMF and other NMF algorithms on several face databases, validate the effectiveness of the proposed method.

Clustering, classification, and regression, are three major research topics in machine learning. So far, much work has been conducted in solving multiple instance classification and multiple instance regression problems, where supervised training patterns are given as bags and each bag consists of some instances. But the research on unsupervised multiple instance clustering is still limited . This paper formulates a novel Maximum Margin Multiple Instance Clustering problem for the multiple instance clustering task. To avoid solving a non-convex  optimization problem directly, M 3 IC is further relaxed, which enables an efficient optimization solution with a combination of Constrained Concave-Convex Procedure CCCP) and the Cutting Plane method. Furthermore, this paper analyzes some important properties of the proposed method and the relationship between the proposed method and some other related ones. An extensive set of empirical results demonstrate the advantages of the proposed method against existing research for both effectiveness and efficiency.

This paper proposes a novel framework for image retrieval. The retrieval is treated as searching for an ordered cycle in an image database. The optimal cycle can be found by minimizing the geometric manifold entropy of images. The minimization is solved by the proposed method, fast active tabu search. Experimental results demonstrate the framework for image retrieval is feasible and quite promising.

Most of the existing metric learning methods are accomplished by exploiting pairwise constraints over the labeled data and frequently suffer from the insufficiency of training examples.  To learn a robust distance metric from few labeled examples, prior knowledge from unlabeled examples as well as the metrics previously derived from auxiliary data sets can be useful.  In this paper, we propose to leverage such auxiliary knowledge to assist distance metric learning, which is formulated following the regularized loss minimization principle.  Two algorithms are derived on the basis of manifold regularization and log-determinant divergence regularization technique, respectively, which can simultaneously exploit label information (i.e., the pairwise constraints over labeled data), unlabeled examples, and the metrics derived from auxiliary data sets.  The proposed methods directly manipulate the auxiliary metrics and require no raw examples from the auxiliary data sets, which make them efficient and flexible.  We conduct extensive evaluations to compare our approaches with a number of competing approaches on face recognition task.  The experimental results show that our approaches can derive reliable distance metrics from limited training examples and thus are superior in terms of accuracy and labeling efforts.

Event detection is a critical task in sensor networks for a variety of real-world applications. Many real-world events often exhibit complex spatio-temporal patterns whereby they manifest themselves via observations over time and space proximities. These spatio-temporal events cannot be handled well by many of the previous approaches. In this paper, we propose a new Spatio-Temporal Event Detection (STED) algorithm in sensor networks based on a dynamic conditional random field (DCRF) model. Our STED method handles the uncertainty of sensor data explicitly and permits neighborhood interactions in both observations and event labels. Experiments on both real data and synthetic data demonstrate that our STED method can provide accurate event detection in near real time even for large-scale sensor networks.

We propose a probabilistic transfer learning model that uses task-level features to control the task mixture selection in a hierarchical Bayesian model. These task-level features, although rarely used in existing approaches, can provide additional information to model complex task distributions and allow effective transfer to new tasks especially when only limited number of data are available. To estimate the model parameters, we develop an empirical Bayes method based on variational approximation techniques. Our experiments on information retrieval show that the proposed model achieves significantly better performance compared with other transfer learning methods.

Due to their flexible nonparametric nature, Gaussian process models are very effective at solving hard machine learning problems. While existing Gaussian process models focus on modeling one single relation, we present a generalized GP model, named multi-relational Gaussian process model, that is able to deal with an arbitrary number of relations in a domain of interest. The proposed model is analyzed in the context of bipartite, directed, and undirected univariate relations. Experimental results on real-world datasets show that exploiting the correlations among different entity types and relations can indeed improve prediction performance.

Feature selection can be conducted in a supervised or unsupervised manner, in terms of whether the label information We consider the problem of semi-supervised feature is utilized to guide the selection of relevant features. Generally, selection, where we are given a small amount supervised feature selection methods require a large of labeled examples and a large amount of unlabeled amount of labeled training data. It however could fail to identify examples. Since a small number of labeled the relevant features that are discriminative to different samples are usually insufficient for identifying the classes, provided the number of labeled samples is small. On relevant features, the critical problem arising from the other hand, while unsupervised feature selection methods semi-supervised feature selection is how to take could work well with unlabeled training data, they ignore advantage of the information underneath the unlabeled the label information and therefore are often unable to identify data. To address this problem, we propose the discriminative features. Given the high cost in manually a novel discriminative semi-supervised feature labeling data, and at the same time abundant unlabeled selection method based on the idea of manifold data are often easily accessible, it is desirable to develop feature regularization. The proposed method selects selection methods that are capable of exploiting both labeled features through maximizing the classification margin and unlabeled data.