This extended abstract details our submission to the Music Information Retrieval Evaluation eXchange (MIREX) 2011 for the audio tag classification task. First of all, we extract a fixed-length feature vector (composed of some timbral as well as modulation spectrum features) from each song clip. Then, by using l-reconstruction to represent each test song clip as a linear combination of all train...

This paper deals with the local learning approach for clustering, which is based on the idea that in a good clustering, the cluster label of each data point can be well predicted based on its neighbors and their cluster labels. We propose a novel local learning based clustering algorithm using kernel regression as the local label predictor. Although sum of absolute error is used instead of sum ...

An important problem in multi-label classification is to capture label patterns or underlying structures that have an impact on such patterns. One way of learning underlying structures over labels is to project both instances and labels into the same space where an instance and its relevant labels tend to have similar representations. In this paper, we present a novel method to learn a joint sp...

Recent advances in language modeling such as word2vec motivate a number of graph embedding approaches by treating random walk sequences as sentences to encode structural proximity in a graph. However, most of the existing principles of neural graph embedding do not incorporate auxiliary information such as node content flexibly. In this paper we take a matrix factorization perspective of graph ...

Extreme multi-label learning (XML) or classification has been a practical and important problem since the boom of big data. The main challenge lies in the exponential label space which involves 2 possible label sets when the label dimension L is very large, e.g., in millions for Wikipedia labels. This paper is motivated to better explore the label space by building and modeling an explicit labe...

Learning with limited supervision presents a major challenge to machine learning systems in practice. Fortunately, various types of extra information exist in real-world problems, characterizing the properties of the model space, the feature space and the label space, respectively. With the goal of supervision reduction, this thesis studies the representation, discovery and incorporation of ext...

Multi-label canonical correlation analysis (ml-CCA) has been developed for cross-modal retrieval. However, the computation of ml-CCA involves dense matrices eigendecomposition, which can be computationally expensive. In addition, only takes semantic into account ignores feature correlation. this paper, we propose a novel framework to simultaneously integrate and By using transformation, show th...

The concept of generating trees has been introduced in the literature by Chung, Graham, Hoggat and Kleiman in [4] to examine Baxter permutations. This technique has been successfully applied by West [17, 18] to other classes of permutations and more recently to some other combinatorial classes such as plane trees and lattice paths (see Barcucci et al. [2]). In all these cases, a generating tree...

Emotional information plays an important role in various multimedia applications. Movies, as a widely available form of content, can induce multiple positive emotions and stimulate people's pursuit better life. Different from negative emotions, are highly correlated difficult to distinguish the emotional space. Since different often induced simultaneously by movies, traditional single-target or...