Sharif Digital Repository

Event extraction is one of the important tasks in Natural Language Processing (NLP). Many NLP applications like question answering, information extraction and summarization need to have some knowledge about events of input documents. There are several definitions for events in NLP domains. In this dissertation, the event is viewed as an element in a network of temporal information. Therefore, the project has been based on the ISO-TimeML specification language, which is the standard scheme for temporal information processing in natural texts. Event extraction based on ISO-TimeML has been performed for a number of languages including English, French, Spanish, and Korean. However, for Persian... 

The purpose of this study is to identify the False Data Injection (FDI) attack and reconstruct the incorrect data created by FDI in the power system using machine learning algorithms. Unlike conventional power grids, the smart grids due to the increase of smart devices and communication networks to transfer power grid information from one point to another and the need to control and monitor the power grid is an electrical network that is integrated with a communication network (cyber-physical system). A communication network that transmits data between the control center and smart meters increases the threat of cyber-attacks in the power grid. In this study, the aim is to investigate the FDI... 

Nowadays, with the increasing size of data,it’s impossible to collect data and fast classification by human, and needs for an automated classification and data analysis, is more interested. Data classification is a process of giving the training data along with their class labels to the learning agent, which learns the relation between the instances and the labels. Then make a prediction to the label of the training data.In this thesis we will observe the classification of the multi-label data. Multi-label data have more than one label. In other words, each instance appears with a vector of labels.In this thesis, a method based on nearest neighbor is proposed to classify the multi-label... 

In this research, an optical processing method for DNA classification is presented in order to overcome the problems in the previous methods. With improving in the operational capacity of the sequencing process, which has increased the number of genomes, comparing sequences with a complete database of genomes is a serious challenge to computational methods. Most current classification programs suffer from either slow classification speeds, large memory requirements, or both. To achieve high speed and accuracy at the same time, we suggest using optical processing methods. The performance of electronic processing-based computing, especially in the case of large data processing, is usually... 

In large databases, availability of labeled training data is mostly prohibitive in classification. Semi-supervised algorithms are employed to tackle the lack of labeled training data problem. Video databases are the epitome for such a scenario; that is why semi-supervised learning has found its niche in it. Graph-based methods are a promising platform for semi-supervised video classification. Based on the multiview characteristic of video data, different features have been proposed (such as SIFT, STIP and MFCC) which can be utilized to build a graph. In this project, we have proposed a new classification method which fuses the results of manifold regularization over different graphs. Our... 

The main objective of this project is to discuss a DOA estimation system performance in a multipath environment. For this goal, first the basics of the DOA Estimation systems are discussed. Evaluation of different types of algorithms and choosing an appropriate one is an important step in this regard. Moreover the comparison of different array structures is necessary to choose the appropriate one. The major part of the project has been dedicated to discussion of multipath phenomenon as a major source of error in DOA estimation and studying the multipath mitigation algorithms to eliminate the aforementioned error. A model based on the concept of Bistatic Radar Cross Section is proposed to... 

Governments and organizations need to classify network trac using deep packet inspection systems, by protocols, applications, and user’s behavior, to monitor, control, and enforce law and governance to the online behavior of its citizens and human resources. The high capacity of machine learning in the classication problem has led trac monitoring systems to use machine learning.The development of machine learning-based trac monitoring systems in the eld of research has reached relative maturity and has reached the border of industrial, commercial and governmental use. In the latest trac classi-cation studies using neural networks, as the most ecient machine learning methods, the classication... 

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... 

In this project we proposed a new method for improving the detection abality of a steganalyser with a pre-processing on contents of an image. Steganalysis, using machine learning, is designing a classifier with two classes: Stego or Cover. This classifier should be trained with extracted features from signal. The result of the training procedure is a machine that decides a signal belongs to stego or cover class. The first step of steganalysis process is extraction of proper features from signal. Proper feature is a variable that represents all of the useful properties of signal. Second step of this process is classifying data to two class of stego and cover. Many algorithms are proposed for... 

Today many applications produce data stream that grow continuously over time. Classification is an important task in mining this data. One of the major challenges in classification of data streams is that the underlying concept of data may change over time; this needs updating the classification model. Many of traditional classification algorithms are adapted for classifying data streams. Evolutionary algorithms are stochastic methods that have been successfully applied to a wide range of optimization problems. Harmony search is a new evolutionary algorithm that convergence to global optimum and can be applied to discrete and continuous data. The evolutionary nature of classifying data... 

By expanding the use of IT and public access to financial markets, the number of players in this area has increased and the nonlinearity of the market has become more complex. Hence, investors need specific strategies that can make profitable investment by determining the time of purchase and sale of stocks. The purpose of this research is to provide a stock trading framework for strategic portfolio management. This framework uses daily values of 18 indicators of technical analysis as features and daily trading signals as data labels for training various machine learning models, such as support vector regression, k nearest neighbors, decision tree, artificial neural network and random... 

Today we live in an era of continuous technology improvement which results in huge changes in different areas of diverse industries. Among the most recent systems for monitoring and quality control which benefits from high speed, are machine vision systems. The output of these systems, are digital images that can be used for monitoring instead of the original products. Unfortunately due to the computational complexity of data extracted from the digital images, traditional methods lose their efficiency. Therefore, in this thesis, a method is proposed to design a model for the monitoring and control of image-based processes, which uses classification methods, that are capable of classifying... 

Today, with the development of technology, cyberattacks are on the rise. Personal and corporate computer systems can be exposed to various threats and dangers of hackers and malware, including information theft, forgery, and denial of service, which can cause great material and moral damage to individuals and organizations. So, it is necessary to take security measures in this regard. Many security mechanisms are available to prevent security vulnerabilities against various threats. In this study, first, after carefully studying network attacks, we identify the criteria for identifying attacks that can be executed in network traffic and explain how to calculate them. The current research... 

Measuring emotions of music is one of the methods to determine music content. Music emotion detection is applicable in music retrieval, recognition of music genre and also music data management softwares. Music emotion is considered in different sciences such as physiology, psychology, musicology and engineering. First, we collected a database of different types of music with various emotions. These data have been labeled according to their emotions. In this project, four emotions (Angry, happy, relax and sad) have been used as labels based on Thayer’s two dimension emotion model. There are two basic steps for music emotion recognition similar to other recognition systems: Feature extraction... 

One of the important jobs of ESM and Elint receiver is to find the direction of arrival of the targets. Also DOA estimation has applications in guiding missiles and jammers location, position finding using of multiple DF systems which are placed in different locations is the most important use of this type of system.
A DF system has multiple parts: receiver, algorithm and arrays of antennas.
Amplitude comparison algorithms are more simple but on the other hand phase comparison algorithms prepare more accurate estimation. In this thesis these algorithms are simulated and to each other. And it will be concluded that, in the application being test, the "MUSIC" algorithm is the best.... 

Cancer is the leading cause of death worldwide, and colorectal cancer is the second leading cause of death in women and the third in men. On the other hand, colon polyps can cause colorectal cancer. Therefore, early detection of polyps is of great importance. In recent years, many methods have been proposed for polyp detection using deep learning with high accuracy, but most of them have problems with speed, accuracy, or interpretability. Speed is important because colonoscopy should be performed as quickly and promptly as possible, and in many cases, it is not possible to repeat the colonoscopy. In addition, many of them only address the issue of polyp detection, while from a medical point... 

In the last two decades, Convolutional Neural Networks (CNN) have shown significant capabilities in artificial intelligence. These networks are able to provide comprehensive conclusions about the overall behavior a system by analyzing the relationship between the components of that system; Clearly, these networks have been successful in performing categorization tasks. However, there are no coherent theories as to why they work, and how to optimize them. On the other hand, according to recent research on the relationship between deep networks (in computer science) and Renormalization Group (in physics), convolutional networks seem to use a method similar to the Density Matrix Renormalization... 

Web services are some kind of software applications which are available on the Web. Growing the popularity of Web services, led to increasing number of providers and as a result a great deal of Web services. This huge number of services made the searching and discovery tasks hard and effort-full jobs. In order to have a better discovery it is better to first classify Web services into some categories and then search in the relevant class. Although this classification can be done based on matching key words through the service registration information, such syntax-level service facilities can’t achieve the satisfaction results both in the precision and the recall sides. Human experience shows... 

Emotion recognition from speech signals has become one of the most popular researches in recent years. In order to increase human-machine interaction, a proper connection must be established between them. To achieve this goal, a machine must be able to understand the situation and respond accordingly. Part of this process involves understanding the user's emotional state. In recent years, various methods have been proposed to increase the efficiency of the speech emotion recognition system. These methods include collecting various audio databases, extracting efficient features from speech signals, using feature selection algorithms, designing different classifiers, as well as combining... 

Distance Metric Learning algorithms have been widely used in Machine Learning methods recently. In these algorithms a distance function between objecs (data points) is learned based on their labels or similarity and dissimilarity constraints. Recent works have shown that a good precision is obtained in classification or clustering methods which use these functions. Since in the current systems many of data points do not exist at the beginning and are added to the training set as the algorithm is run, online methods are needed to update learned metric due to new data.
In this thesis, we proposed a new online distance metric learning method that has higher performance than existing...