Sharif Digital Repository

In some of object recognition problems, labeled data may not be available for all categories. Zero-shot learning utilizes auxiliary information (also called signatures) describing each category in order to find a classifier that can recognize samples from categories with no labeled instance. On the other hand, with recent advances made by deep neural networks in computer vision, a rich representation can be obtained from images that discriminates different categorizes and therefore obtaining a unsupervised information from images is made possible. However, in the previous works, little attention has been paid to using such unsupervised information for the task of zero-shot learning. In this... 

Large data sources labeled by experts at cost are essential for deep learning success in various domains. But, when labeling is expensive and labeled data is scarce, deep learning generally does not perform well. The goal of semi-supervised learning is to leverage abundant unlabeled data that one can easily collect. New semi-supervised algorithms based on data augmentation techniques have reached new advances in this field. In this work, by studying different textual augmentation techniques, a new approach is proposed that can obtain effective information signals from unlabeled data. The method encourages the model to generate the same representation vectors for different augmented versions... 

Recent advances in storage and processing, have provided the ability of automatic gathering of information which in turn leads to fast and contineous flow of data. The data which are produced and stored in this way, are named data streams. data streams have many applications such as processing financial transactions, the recorded data of various sensors or the collected data by web sevices. Data streams are produced with high speed, large size and much dynamism and have some unique properties which make them applicable in precise modeling of many real data mining applications. The main challenge of data streams is the occurrence of concept drift which can be in four types: sudden, gradual,... 

Classification of hyperspectral remote sensing images is a challenging problem, because of the small number of labeled pixels, high dimensionality of the data and large number of pixels. In this context, semisupervised learning can improve the classification accuracy by extracting information form the distribution of all the labeled and unlabeled data. Among semi-supervised methods, manifold-based algorithms have been frequently used in recent years. In most of the previous works, manifolds are constructed according to spectral representation of data, while spatial dependency of pixel labels is an important property of the images in remote sensing applications. In this thesis, after... 

Image classification, i.e. the task of assigning an image to a class chosen from a predefined set of classes, has addressed in this thesis. At first the classifier is divided into two major sub partitions, feature extraction and classifier. Then we show that by using local feature extraction techniques such as BOW the classification accuracy will improve. In addition, using unlabeled data is argued as the fact to deal with high nonlinear structure of features. 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. So in large... 

Many Text-to-Speech (TTS) systems, particularly in low-resource environments, struggle to produce natural and intelligible speech from grapheme sequences. One solution to this problem is to use Grapheme-to-Phoneme (G2P) conversion to increase the information in the input sequence and improve the TTS output. However, current G2P systems are not accurate or efficient enough for Persian texts due to the language’s complexity and the lack of short vowels in Persian grapheme sequences. In our study, we aimed to improve resources for the Persian language. To achieve this, we introduced two new G2P training datasets, one manually-labeled and the other machine-generated, containing over five million...