Sharif Digital Repository

Image annotation approaches need an annotated dataset to learn a model for the relation between images and words. Unfortunately, preparing a labeled dataset is highly time consuming and expensive. In this work, we describe the development of an annotation system in semi-supervised learning framework which by incorporating unlabeled images into training phase reduces the system demand to labeled images. Our approach constructs a generative model for each semantic class in two main steps. First, based on Gamma distribution, a generative model is constructed for each semantic class using labeled images in that class. The second step incorporates the unlabeled images by using a modified EM... 

Recent advances in storage and processing have provided the possibility of automatic gathering of information, which in turn leads to fast and continuous flows of data. The data which are produced and stored in this way are called data streams. Data streams are produced in large size, and much dynamism and have some unique properties which make them applicable to model many real data mining applications. The main challenge of streaming data is the occurrence of concept drift. In addition, regarding the costs of labeling of instances, it is often assumed that only a small fraction of instances are labeled. In this paper, we propose an ensemble algorithm to classify instances of non-stationary... 

In this paper, we propose two new algorithms for transductive multi-label learning from missing data. In transductive matrix completion (MC), the challenge is prediction while the data matrix is partially observed. The joint MC and prediction tasks are addressed simultaneously to enhance accuracy in comparison with separate tackling of each. In this setting, the labels to be predicted are modeled as missing entries inside a stacked matrix along the feature-instance data. Assuming the data matrix is of low rank, we propose a new recommendation method for transductive MC by posing the problem as a minimization of the smoothed rank function with non-affine constraints, rather than its convex... 

Multi-label classification has received considerable interest in recent years. Multi-label classifiers usually need to address many issues including: handling large-scale datasets with many instances and a large set of labels, compensating missing label assignments in the training set, considering correlations between labels, as well as exploiting unlabeled data to improve prediction performance. To tackle datasets with a large set of labels, embedding-based methods represent the label assignments in a low-dimensional space. Many state-of-the-art embedding-based methods use a linear dimensionality reduction to map the label assignments to a low-dimensional space. However, by doing so, these... 

Market prediction has been an important research problem for decades. Having better predictive models that are both more accurate and faster has been attractive for both researchers and traders. Among many approaches, semi-supervised graph-based prediction has been used as a solution in recent researches. Based on this approach, we present two prediction models. In the first model, a new network structure is introduced that can capture more information about markets’ direction of movements compared to the previous state of the art methods. Based on this novel network, a new algorithm for semi-supervised label propagation is designed that is able to prediction the direction of movement faster... 

When the number of labeled data is not sufficient, Semi-Supervised Learning (SSL) methods utilize unlabeled data to enhance classification. Recently, many SSL methods have been developed based on the manifold assumption in a batch mode. However, when data arrive sequentially and in large quantities, both computation and storage limitations become a bottleneck. In this paper, we present a new semi-supervised coarse graining (CG) algorithm to reduce the required number of data points for preserving the manifold structure. First, an equivalent formulation of Label Propagation (LP) is derived. Then a novel spectral view of the Harmonic Solution (HS) is proposed. Finally an algorithm to reduce... 

Almost all of the existing domain adaptation methods assume that all test data belong to a single stationary target distribution. However, in many real world applications, data arrive sequentially and the data distribution is continuously evolving. In this paper, we tackle the problem of adaptation to a continuously evolving target domain that has been recently introduced. We assume that the available data for the source domain are labeled but the examples of the target domain can be unlabeled and arrive sequentially. Moreover, the distribution of the target domain can evolve continuously over time. We propose the Evolving Domain Adaptation (EDA) method that first finds a new feature space... 

Graph based methods are among the most active and applicable approaches studied in semi-supervised learning. The problem of neighborhood graph construction for these methods is addressed in this paper. Neighborhood graph construction plays a key role in the quality of the classification in graph based methods. Several unsupervised graph construction methods have been proposed that have addressed issues such as data noise, geometrical properties of the underlying manifold and graph hyper-parameters selection. In contrast, in order to adapt the graph construction to the given classification task, many of the recent graph construction methods take advantage of the data labels. However, these...