Loading...
Search for:
learning
0.025 seconds
Total 1677 records
Persistent Homology and its Applications in Machine Learning
, M.Sc. Thesis Sharif University of Technology ; Ranjbar, Alireza (Supervisor)
Abstract
Persistent homology is one of the main tools in Topological Data Analysis. Indeed, to deal with a huge dataset while noise sensitivity is important, persistent homology can reflect some information about data in the form of persistent homology groups and persistence diagrams. Note that statistical or linear algebraic tools are not suitable to work with huge datasets with very high dimensions. In this thesis, we discuss the concept of persistent homology and investigate some of its properties such as the stability of the persistence diagrams. Indeed, persistence diagrams are obtained from the generating sets of the persistent homology groups. Further, we discuss an application of persistent...
Designing an Automatic System for Continuous Meaningful Gesture Recognition by Deep Learning
, M.Sc. Thesis Sharif University of Technology ; Taheri, Alireza (Supervisor)
Abstract
Meaningful gesture recognition, whose purpose is to interpret human movements, plays a crucial role in various fields such as human and computer interaction, sign language recognition, robot control, and medical applications. Sign language recognition is regarded as the most significant use of gesture recognition by many researchers. Sign languages are the natural medium of communication for millions of deaf people all over the world, and the existence of a sign language recognition system has significantly aided in facilitating communication between deaf individuals and others. Despite numerous studies conducted in this field in recent years, there are still many challenges to continuous...
Learning of Causal Structures with Deep Reinforcement Learning
, M.Sc. Thesis Sharif University of Technology ; Saleh Kaleybar, Saber (Supervisor) ; Hashemi, Matin (Co-Supervisor)
Abstract
We study the problem of experiment design to learn causal structures from interventional data. We consider an active learning setting in which the experimenter decides to intervene on one of the variables in the system in each step and uses the results of the intervention to recover further causal relationships among the variables. The goal is to fully identify the causal structures with minimum number of interventions. We present the first deep reinforcement learning based solution for the problem of experiment design. In the proposed method, we embed input graphs to vectors using a graph neural network and feed them to another neural network which outputs a variable for performing...
Detecting Metastatic Lung Cancer and Its Lesions From CT-Scan Images Using Deep Interpretable Networks
, M.Sc. Thesis Sharif University of Technology ; Rabiee, Hamid Reza (Supervisor)
Abstract
Using automated assistants in medical applications has been increased in recent years. One of the most popular methods are artificial intelligence and deep learning methods which are specifically used in medical images analysis. Using these methods can improve the diagnosis accuracy, while performing in a faster time. So these methods can reduce the economical costs, error rate, and response time. But one important challenge in deep learning methods, is the interpretability of neural networks. In this research we focused on introducing an interpretability method for our pixel-wise segmentation network which is applied to the lung nodules dataset. In this research we first implemented a...
Identification and Forecasting of Nuclear Power Plants Transients by Semi-Supervised Method with Change of Representation Technique
, M.Sc. Thesis Sharif University of Technology ; Ghofrani, Mohamad Bagher (Supervisor) ; Moshkbar Bakhshayesh, Khalil (Supervisor)
Abstract
In this work, we aim to find a way to identify and forecast transients in nuclear power plants with the aid of semi-supervised machine learning algorithm. Forecasting and identifying transients in nuclear power plants at the early stages of formation are essential for safety considerations and precautionary measures. The use of machine learning algorithms provides an intelligent control mechanism that, along with the main operator of the power plant, raises the transient detection and identification rate. Our algorithm of choice is to change the way data is presented, which is a semi-supervised learning approach. The algorithm consists of two methods: quantum dynamics clustering...
A Study in Genome Editing with Clustered Regularly Interspaced Short Palindromic Repeats
, M.Sc. Thesis Sharif University of Technology ; Sharifi Tabar, Mohsen (Supervisor) ; Rabiee, Hamid Reza (Co-Supervisor) ; Rohban, Mohammad Hossein (Co-Supervisor)
Abstract
Clustered Regularly Interspaced Short Palindromic Repeats, or in short, CRISPR is a relatively new technology that enables geneticists and medical researchers to edit parts of the genome by removing, adding, or altering parts of the DNA. Initially found in the genomes of prokaryotic organisms such as bacteria and archaea, this technology can cure many illnesses such as blindness and cancer. A significant issue for a practical application of CRISPR systems is accurately predicting the single guide RNA (sgRNA) on-target efficacy and off-target sensitivity. While some methods classify these designs, most algorithms are on separate data with different genes and cells. The lack of...
Persian Query Corrector Based on Deep Learning Networks (with Emphasis on Spatial Queries)
, M.Sc. Thesis Sharif University of Technology ; Izadi, Mohammad (Supervisor)
Abstract
The main spectrum of a location search engine is retrieving the locations that are most relevant to the user query. Geographic and spatial queries usually consist of a series of keywords that express the user's needs. The geographic search engine should retrieve the locations associated with the user request and then rank the retrieved results according to their relevance to the user request. In some cases, user requests may contain spelling errors that can greatly affect the results retrieved. Spell correction is automatic in the back spectrum recovery system. Spelling correction is the task of automatically recovering the intended text from a misspelled text. Therefore, spelling correction...
Causal Discovery and Generative Neural Networks to Identify the Functional Causal Model
, M.Sc. Thesis Sharif University of Technology ; Bahraini, Alireza (Supervisor)
Abstract
Causal discovery is of utmost importance for agents who must plan and decide based on observations. Since mistaking correlation with causation might lead to un- wanted consequences. The gold standard to discover causal relation is to perform experiments. However, experiments are in many cases expensive, unethical or impossible to perform. In these situations, there is a need for observational causal discovery. Causal discovery in the observational data setting involves making significant assumptions on the data and on the underlying causal model. This thesis aims to alleviate some of the assumptions and tries to identify the causal relationships and causal mechanisms using generative neural...
A Deep Generative Model for Graph-Structured Data
, M.Sc. Thesis Sharif University of Technology ; Movaghar, Ali (Supervisor)
Abstract
In recent years, deep generative models have achieved incredible successes in various fields, including graph generation. Due to the advances made in graph generation by deep generative models, these methods have shown numerous applications from drug discovery and molecular graph generation to modeling social and citation network graphs. Graph generation is an approach to discovering and exploring new graph structures and has been attracting growing attention. One of the most challenging applications of deep graph generative models is molecular graph generation since it requires not only generating chemically valid molecular structures but also optimizing their chemical properties in the...
Few-Shot Semantic Segmentaion Using Meta-Learning
,
M.Sc. Thesis
Sharif University of Technology
;
Soleymani Baghshah, Mahdieh
(Supervisor)
Abstract
Despite recent advancements in deep learning methods, these methods rely on a huge amount of training data to work. Recently the problem of solving classification and recently semantic segmentation problems with a few training data have gained attention to tackle this issue. In this research, we propose a meta-learning method by combining optimization-based and prototypical approaches in which a small portion of parameters are optimized with task-specific initialization. In addition to this and designing other parts of the method, we propose a new approach to use query data as an unlabeled sample to enhance task-specific learning. Alongside the mentioned method, we propose an approach to use...
Graph Learning from Incomplete and Noisy Graph Signals
, M.Sc. Thesis Sharif University of Technology ; Babaiezadeh, Masoud (Supervisor)
Abstract
The problem of inferring a graph from a set of graph signals over it plays a crucial role in the field of Graph Signal Processing (GSP). When provided with a graph that best models the structure of data, the GSP algorithms can offer high data processing capability. However, a meaningful graph of data is not always available, hence in some applications, the graph needs to be learned from the data itself. When the data is corrupted and missing, this task becomes even more challenging. In this paper, we present a graph learning algorithm that is capable of learning the underlying structure of data from an incomplete and noisy dataset of graph signals. We propose an algorithm that jointly...
A Novel Resource Allocation Algorithm in Edge Computing with Deep Reinforcement Learning
, M.Sc. Thesis Sharif University of Technology ; Movaghar, Ali (Supervisor)
Abstract
With the explosion of mobile smart devices, many computation intensive applications have emerged, such as interactive gaming and augmented reality. Mobile edge computing (EC) is put forward, as an extension of cloud computing, to meet the low-latency require- ments of the applications. In mobile edge computing systems, an edge node may have a high load when a large number of mobile devices offload their tasks to it. those offloaded tasks may experience large processing delay or even be dropped when their deadlines expire. Due to the uncertain load dynamics at the edge nodes, it is challenging for each device to determine its offloading decision (i.e., whether to offload or not, and which...
“Xylotism”: A tablet-based application to teach music to children with autism
, Article 9th International Conference on Social Robotics, ICSR 2017, 22 November 2017 through 24 November 2017 ; Volume 10652 LNAI , 2017 , Pages 728-738 ; 03029743 (ISSN); 9783319700212 (ISBN) ; Habibnejad Korayem, A ; Shariati, A ; Meghdari, A ; Alemi, M ; Ahmadi, E ; Taheri, A ; Heidari, R ; Sharif University of Technology
Springer Verlag
2017
Abstract
Technology is inevitable, and its role for clinical therapists and specialists cannot be ignored. The promising movement towards computer-based interventions, specifically the use of tablets as an effective and newly developed learning device for children with autism spectral disorder (ASD) highlights the role of technology in addressing the shortcomings of conventional therapy methods. In this paper, we present a new application, named as Xylotism, which is an interactive game to improve learning and teach music to children with autism spectrum disorder. The game can be played with/without parents/therapists’ involvement, which increases its usefulness and effectiveness. We have...
A Lattice-based Authenticated Group Key Establishment Scheme Using Secret Sharing
, M.Sc. Thesis Sharif University of Technology ; Aref, Mohammad Reza (Supervisor) ; Eghlidos, Taraneh (Co-Supervisor)
Abstract
Secure communication among members of a group requires a shared cryptographic key. To address this issue, group key agreement and exchange schemes are introduced. In group key exchange schemes, a trusted center generates a shared key for the group and sends it securely to the group members. But in group key agreement schemes, all members of the group are involved in generating group keys. One of the useful ways in these schemes is to use secret sharing to share keys. A public key infrastructure is used to secure the distribution of the shares in secret sharing schemes. As quantum computers threat the classic cryptographic algorithms that are based on the difficulty of factoring large numbers...
A Persian Dialog System with Sequence to Sequence Learning
, M.Sc. Thesis Sharif University of Technology ; Sameti, Hossein (Supervisor)
Abstract
Conversation modeling is one of the most important goals in the field of understanding natural language and machine intelligence. Recently, with the enormous growth of the Internet and social networks, the amount of available data on the Web has increased significantly.This makes it possible to use data-driven approaches to solve the modeling problem of conversation.One of the most recent data-driven methods is the sequence to sequence modeling. In this document, after providing the necessary prerequisites, we examined the various models that have used the sequence to sequence approach for conversation modeling. We further examined the ways of improving the efficiency of this modeling...
A Machine Learning Approach to Minimize Power Consumption of Smartphones While Satisfying the Gaming Performance
, M.Sc. Thesis Sharif University of Technology ; Sarbazi Azad, Hamid (Supervisor)
Abstract
Today's smartphone devices include several cores, such as CPU, GPU, and different accelerators, in order to maximize user experience. However, due to meeting the power budget and limited capacity battery, power and energy of their cores should be managed using dynamic power management methods such as dynamic voltage and frequency scaling (DVFS). For this purpose, we should find optimal frequency and voltage settings of processing cores for each time, to minimize energy consumption while retaining user experience. Finding this optimal frequency and voltage settings is a challenging problem that depends on many parameters. We propose to use deep reinforcement learning (DRL) method to...
A Systematic Approach for Biomarker Identification in Autism Spectrum Disorder based on Machine learning
, M.Sc. Thesis Sharif University of Technology ; Jafari Siavoshani, Mahdi (Supervisor) ; Kavousi, Kaveh (Co-Supervisor) ; Ohadi, Mina (Co-Supervisor)
Abstract
Autism spectrum disorder (ASD) is a strong genetic perturbation that encompasses a wide range of clinical symptoms, including functional at different regions of the brain, repetitive behaviors, and interests, weaknesses in social relationships, some sensitivities to environmental factors and etc. Genetic complexity and the impact of environmental factors put the disease in the category of Level 1 complex developmental disorders.We proposed a pilot, combined, and highly effective structure to identify biomarkers in the autism spectrum disorder that could be extended to other diseases that have a similar genetic architecture with autism. We also develop a Gene-tissue interaction network to...
A Machine Learning-Based Hierarchical Risk Parity Approach for Portfolio Asset Allocation on the Tehran Stock Exchange
, M.Sc. Thesis Sharif University of Technology ; Habibi, Moslem (Supervisor) ; Fazli, Mohammad Amin (Co-Supervisor)
Abstract
The process of portfolio construction and optimization can be broken down into three main steps: selecting appropriate assets, allocating capital, and monitoring and adjusting the portfolio. This study focuses on evaluating the performance of the Hierarchical Risk Parity (HRP) method for capital allocation in investment portfolios, specifically in Iran’s capital market. The aim is to enhance the method's effectiveness by implementing alternative correlation calculation approaches, such as Wavelet and Chatterjee correlations. The study utilizes three different portfolios containing assets from the Tehran Stock Exchange, the US stock market, and the cryptocurrency market. The primary objective...
A Deep Learning Approach to Classify Motor Imagery Based on The Combination of Discrete Wavelet Transform and Convolutional Neural Network for Brain Computer Interface System
, M.Sc. Thesis Sharif University of Technology ; Selk Ghafari, Ali (Supervisor) ; Zabihollah, Abolghssem (Supervisor)
Abstract
A Brain-Computer Interface (BCI) is a communication system that does not need any peripheral muscular activity. The huge goal of BCI is to translate brain activity into a command for a computer. One of the most important topics in the brain-computer interface is motor imagery (MI), which shows the reconstruction of subjects. The electrical activities of the brain are measured as electroencephalogram (EEG). EEG signals behave as low to noise ratio also show the dynamic behaviors.In the present work, a novel approach has been employed which is based on feature extraction with discretion wavelet transform (DWT), support vector machine (SVM), Artificial Neural Network (ANN) and Convolutional...
Developing a Multi-Skilled Project Scheduling Problem Model Considering Costs
, M.Sc. Thesis Sharif University of Technology ; Shadrokh, Shahram (Supervisor)
Abstract
This research intends to investigate the multi-skilled project scheduling problems (MSPSP) considering staffs’ dynamic efficiency. In doing so, a mixed-integer nonlinear programming (MINLP) model is proposed, considering the influence of personnel learning and forgetting on activities duration. The more the staff spends time on their various skills, the more learned and efficient they will be on that skill (according to personal learning curves) so that they can complete new tasks less costly and more quickly. Most project-based organizations want to maximize value by finishing the projects efficiently and developing their staff’s competence through projects. Inherently, any...