PhD courses
Speciality Code:
8D06103
Speciality Name:
Computer Science
Faculty:
Information technology
Qualification:
- Scientific and pedagogical direction - Doctor of Philosophy (PhD)
- Model of graduating student
- Mandatory disciplines
- Elective disciplines
- Professional
ON1 Interpreting fundamental concepts in computer science and new programming paradigms, applying them in software design and development.
ON2 Formulating scientific goals, plan research and conducting large-scale computational experiments in specific applications.
ON3 Critically analyzing, evaluating and synthesizing new and complex ideas in the field of computer science.
ON4 Applying big data processing and data mining methods to solve resource-intensive tasks.
ON5 Developing computational algorithms for engineering tasks and implementing them in high-performance systems.
ON6 Exploring computational complexity and stability of algorithms.
ON7 Analyzing and evaluating the reliability and fault tolerance of computer systems.
ON8 Comparing, analyzing and interpreting complex experimental data and draw conclusions.
ON9 Presenting cutting-edge topics and research results at international and national conferences, seminars and workshops both in front of specialists and in an audience that does not have relevant professional training.
ON10 Contributing to the original research that expands the boundaries of knowledge by developing a significant amount of work, publish the research results in the form of scientific articles in Kazakhstan and foreign publications.
ON11 Compiling explanatory notes and applications for research projects, carry out planning, as well as guide and manage research in computer science and related interdisciplinary areas.
ON12 Organizing research, design, educational and professional activities, participating in scientific, state and industrial research as part of a team, being prepared for correct and tolerant interaction in society, for social interaction and cooperation to solve scientific and technical problems.
ON2 Formulating scientific goals, plan research and conducting large-scale computational experiments in specific applications.
ON3 Critically analyzing, evaluating and synthesizing new and complex ideas in the field of computer science.
ON4 Applying big data processing and data mining methods to solve resource-intensive tasks.
ON5 Developing computational algorithms for engineering tasks and implementing them in high-performance systems.
ON6 Exploring computational complexity and stability of algorithms.
ON7 Analyzing and evaluating the reliability and fault tolerance of computer systems.
ON8 Comparing, analyzing and interpreting complex experimental data and draw conclusions.
ON9 Presenting cutting-edge topics and research results at international and national conferences, seminars and workshops both in front of specialists and in an audience that does not have relevant professional training.
ON10 Contributing to the original research that expands the boundaries of knowledge by developing a significant amount of work, publish the research results in the form of scientific articles in Kazakhstan and foreign publications.
ON11 Compiling explanatory notes and applications for research projects, carry out planning, as well as guide and manage research in computer science and related interdisciplinary areas.
ON12 Organizing research, design, educational and professional activities, participating in scientific, state and industrial research as part of a team, being prepared for correct and tolerant interaction in society, for social interaction and cooperation to solve scientific and technical problems.
-
Research and analysis of algorithms
- Number of credits: 5
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline:formation of the ability to apply the tools and methods necessary to offer algorithmic solutions to real problems that have strict theoretical limitations on the use of time and space. As a result of studying the discipline, the following abilities of doctoral students will be formed: - compare, collate and apply key data structures: trees, lists, stacks, queues, hash tables and graph representation; - analyze algorithms and evaluate their behavior in the worst and average cases (in simple cases); - theoretically compare and analyze the temporal complexity of algorithms and data structures; - make the description of algorithms both in functional, and in procedural styles; - develop new and apply the existing fundamental algorithms and data structures to solve real problems. Within the discipline the following aspects will be considered: Divide and Conquer. The method of recurrence relations. Finding the median for linear time. Search graphs. Dijkstra's algorithm. Connectivity in directed graphs. Introduction to the greedy algorithm. Minimum spanning trees. The Kruskal algorithms and the system of disjoint sets. Path compression and clustering. Introduction to the randomized algorithm. Quick sort. Review of approaches based on probabilities. Hashing Balanced search trees and lists with gaps. NP-complete tasks.
-
Academic Writing
- Number of credits: 2
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: Aim оf discipline: Formation of the skills of doctoral students to present the obtained fundamental and applied results in the relevant field by means of scientific analysis and modern information and communication technologies in the form of scientific and technical written works in accordance with the requirements of leading peer-reviewed international publications with a non-zero impact factor included in the Clarivate databases Analytics (Web of Science Core Collection) and Scopus. Abstract оf discipline: Planning and defining the structure of a scientific article. Preparation and publication of articles in peer-reviewed journals. Academicism of presentation. Title, keywords, resume. Citation. Reviews, reviews and critical reviews. Plagiarism. List of scientific publications recommended for publication of the main results of scientific activity. Scopus database. Database "Web of Knowledge". Using databases for literary searches. Indicators of the effectiveness of publications. Authors' citation index (Hirsch index). Impact Factor of Journal Citation Reports (JCR) on the Web of Science database. SCIMago Journal Rank (SJR). Scopus percentile (CiteScore).
-
Advanced algorithms and complexity
- Number of credits: 5
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline: formation of the ability to perform analysis of algorithms and computational complexity in various areas of the theory of computation. As a result of studying the discipline, the following abilities of doctoral students will be formed: - perceive a weakly defined task and present it as a clearly defined task specification; - evaluate a variety of models of limiting resources, such as information theory, space complexity, parallel complexity, communication complexity, proof complexity, query complexity, and complexity of approximation; - apply various advanced algorithmic methods and evidence methods; - recognize deficiencies in poorly formed evidence; - evaluate the correlation of algorithms and / or computational complexity with various indicators of complexity, such as time, space, communication, or informational content. Within the discipline the following aspects will be considered: Greedy algorithms. Dynamic programming: weighted interval planning, segmented least squares. Dynamic programming: alignment of the sequence, the shortest path on the graph. Network streams. NP-completeness: reduced complexity, the traveling salesman problem. NP-completeness: the sum of subsets, other classes of complexity. Computability: diagonalization and the problem of stopping, reduction to the problem of stopping, and Rice's theorem. Computability: Turing machines and the Church-Turing hypothesis. Approximation algorithms. Parallel computations. Online algorithms.
-
Scientific Research Methods
- Number of credits: 3
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of this discipline is to master the basics of the methodology of scientific research, consideration of different levels of scientific knowledge. Study of the stages of conducting research, including the selection of the direction of research, staging of scientific and technical problems, conducting theoretical and experimental research, recommendations for formalization of the formulation The course is also aimed at reviewing the basics of inventive work, patent search and sample plan for a PhD dissertation
-
PhD thesis writing and defence
- Number of credits: 12
- Type of control: Докторская диссертация
- Description: The main purpose of "PhD thesis writing and defence": of a doctoral dissertation is the formation of the doctoral students' ability to disclose the content of research work for the defense of the thesis. During the study of course, doctoral student's should be competent in: 1. to substantiate the content of new scientifically grounded theoretical and experimental results that allow to solve a theoretical or applied problem or are a major achievement in the development of specific scientific directions; 2. explain the assessment of the completeness of the solutions to the tasks assigned, according to the specifics of the professional sphere of activity; 3. they can analyze alternative solutions for solving research and practical problems and assess the prospects for implementing these options; 4. apply the skills of writing scientific texts and presenting them in the form of scientific publications and presentations. 5. to plan and structure the scientific search, clearly highlight the research problem, develop a plan / program and methods for its study, formalize, in accordance with the requirements of the State Educational Establishment, the scientific and qualification work in the form of a thesis for a scientific degree Doctor of Doctor of Philosophy (PhD) on specialty «8D07502 – Standardization and certification (by industry)». During the study of the discipline doctoral student will learn following aspects: Registration of documents for presentation of the thesis for defense. Information card of the dissertation and registration-registration card (in the format Visio 2003). Extract from the minutes of the meeting of the institution, in which the preliminary defense of the thesis took place. Cover letter to the Higher Attestation Commission. Expert conclusion on the possibility of publishing the author's abstract. Expert opinion on the possibility of publishing a dissertation. Minutes of the meeting of the counting commission. Bulletin for voting. A shorthand record of the meeting of the dissertational council. List of scientific papers. Response of the official opponent. A review of the leading organization. The recall of the scientific adviser.
-
Big data Analytics
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline:formation of the ability to evaluate big data analysis technologies for various use cases and develop own software products. As a result of studying the discipline, the following abilities of doctoral students will be formed: - determine the characteristics of the data sets and compare trivial data and big data for different applications; - solve problems associated with such characteristics of big data as high dimensionality, dynamically growing data and scalability problems; - integrate machine learning libraries and mathematical and statistical tools with modern technologies; - recognize and implement various ways of choosing the appropriate model parameters for different machine learning methods; - develop applications using the neural network apparatus and the Tensor Flow framework. Within the discipline the following aspects will be considered: Big Data Review. Using big data in business. Big data processing technologies. Basic Statistics and R. Relationships and Representations, Graph databases. Introduction to Spark 2.0. Language processing using Spark 2.0. Analysis of streaming data using Spark 2.0. Basic Neural Network and Tensor Flow. Evaluation of the quality of big data analysis. Image analysis, text recognition applications. Speech analysis. Question-answer systems. Analyze streaming data using TensorFlow, VoltDB, DataFlowEngines and other databases.
-
High-performance programming with multi-core and graphics processors
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline: formation of the ability to analyze ways of achieving the potential performance of program execution through the knowledge of the underlying computing platform and its interaction with programs. As a result of studying the discipline, the following abilities of doctoral students will be formed: - describe various modern high-performance processors, in particular, the newest Intel cores and memory hierarchy, multi-core cache and graphics processors; - use extended command sets with implicit code and built-in functions; - program complex equipment to obtain high load; - develop software for modern high-performance processors; - evaluate programming methods on multi-core and graphics processors using applications such as matrix operations and fast Fourier transform. Within the discipline the following aspects will be considered: High-performance computing concepts. Concurrency levels Models of parallel computing. HPC architecture. Concurrent programming with CUDA. Programming models in high-performance computing architectures. Memory hierarchy and memory design for a specific transaction. The main problems of design in parallel computing. Mapping of parallel algorithms to parallel architectures, analysis of the performance of parallel algorithms. The main constraints facing parallel computing. Energy saving communication. Quantum computers.
-
Computational algorithms of engineering problems of hydrodynamics on high-performance systems
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline: formation the ability to solve the Navier-Stokes and Euler equations for engineering problems using computational algorithms and programming on high-performance systems. As a result of studying the discipline, the following abilities of doctoral students will be formed: - describe the mathematical characteristics of partial differential equations; - determine the main properties of computational methods - accuracy, stability, consistency; - computationally solve the Euler and Navier-Stokes equations; - apply the methods of parallelization of problems of hydrodynamics on high-performance systems; - develop software systems for solving engineering problems of hydrodynamics on high-performance systems. Within the discipline the following aspects will be considered: Mathematical tools for analytic descriptions of a mathematical model. Linear systems of equations on parallel computing using distributed memory machines and related software standards. Parallel linear algebra. Derivation of equations governing fluid flow. Equations for incompressible flow and boundary conditions. Finite-difference approximations. Navier Stokes solution for compressible flows. Solution of the Navier-Stokes equations for incompressible flows.
-
Deep learning
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline: formation of the ability to use the available training tools using deep neural networks and develop new ones for analyzing big data. As a result of studying the discipline, the following abilities of doctoral students will be formed: - describe the fundamental principles, theories and approaches to learning with the help of deep neural networks, the main options for deep learning and their typical applications; - use key concepts, problems and practices in training and modeling with deep architectures; - connect the concepts and methods presented in the course with their own research; - implement, train and evaluate neural networks using existing software libraries; - present and critically evaluate current research on neural networks and their applications. Within the discipline the following aspects will be considered: Fundamentals of neural networks and deep learning. Methods for improving neural networks: regularization and optimization. Setting hyper parameters and deep learning framework. Tensorflow. Adjustment of hyperparameters and deep learning frameworks. Keras. Strategies for organizing and successfully building a machine learning project. Convolutional neural networks, their applications. Object classification and related methods. Convolutional neural networks, their applications. Recurrent neural networks, their applications. Natural language processing. Speech recognition and related methods.
-
High Performance Computing Models
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline:formation of the ability to manage the technology of the organization of parallel computing on multiprocessor computing systems with distributed or common RAM. As a result of studying the discipline, the following abilities of doctoral students will be formed: consider the methods of logical representation of the structure of multiprocessor computing systems; analyze the existing computational schemes and implement their decomposition; analyze the complexity of basic data transfer operations; simulate parallel programs; form models of computing systems. Within the discipline the following aspects will be considered: General characteristics of data transfer mechanisms. Analysis of the complexity of basic data transfer operations. Generalized data transfer from one processor to all other network processors. Generalized data transfer from all processors to all network processors.Methods of logical representation of the topology of the communication environment. Evaluation of the complexity of data transfer operations for cluster systems. Simulation of parallel programs. Methodology for developing parallel algorithms. The distribution of subtasks between processors must be performed in such a way that the availability of information links. Methods for solving partial differential equations. The organization of parallel computing for systems with shared memory. Formation of a computer system model. Formulation of a computational problem and the choice of a parallel solution method. Definition of graphical forms of observation of the parallel computing process.
-
Information resource modeling
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline: formation of the ability to model information resources with the methodology of structural and system analysis of information processes and systems. As a result of studying the discipline, the following abilities of doctoral students will be formed: - carry out the development and research of theoretical and experimental models of information resources; - conduct the development and study of methods for analyzing, synthesizing, optimizing and predicting the quality of the processes of functioning of information systems and technologies; - carry out the formulation and conduct of experiments according to a given methodology and analysis of the results; - analyze the results of experiments, select the best solutions, prepare and compile reviews, reports and scientific publications; - predict the development of information systems and technologies. Within the discipline the following aspects will be considered: Current state and general characteristics. Problems of modeling complex systems. Methodology for constructing mathematical models of information processes and systems. System analysis of information processes and systems. Principles of a systematic approach. Typical mathematical modeling schemes. Formalization and algorithmization of systems functioning processes. Methods of mathematical modeling of computer systems. Simulation of random effects on the system. Planning machine experiments with system models. Modeling the functioning of systems in some subject areas. Trends and prospects for the development of research methods and modeling of information processes and technologies. Classification of types of modeling. Methods and stages of development of models of systems. Construction of conceptual models and their formalization.
-
Reliability in distributed systems
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline:formation of the ability to apply approaches, principles and methods to the creation of reliable algorithmic, technical and software for distributed computing systems. As a result of studying the discipline, the following abilities of doctoral students will be formed: own methods, models and modern research tools to assess and ensure the reliability of distributed systems; perform reliability calculations, predict the operation of distributed systems from the point of view of ensuring the specified reliability; develop models and methods to create an effective.; analyze and evaluate the factors affecting the reliability and fault tolerance of distributed systems; implement methods, models and tools in the process of creating effective software packages. Within the discipline the following aspects will be considered: The problem of reliability in the modern world. Reliability terminology. Quantitative indicators of reliability. Continuous and discrete laws of reliability. Related distributions. Accelerated reliability tests. Calculation of reliability in terms of incomplete information. Bayesian approach. Reliability of non-recoverable systems. Calculation using logical algebra. Redundancy as a method of improving reliability. The choice of elements and schemes. Optimization of the cost of reserved objects. Reliability of the restored systems. Calculation of availability factors. Preventive measures as a means of maintaining a given level of reliability. Types of prevention. The choice of optimal strategies and schedules of PPR. Evaluation of the optimal service life of the main equipment. Deterministic and stochastic approach. Markov processes and dynamic programming. The effectiveness of ideal and real preventive. Maximizing equipment efficiency. Software reliability assessment. Control, completion and acceptance of programs. Provision of spare elements in the conditions of external supply and in the presence of a repair base. Organizational issues of reliability.
-
Advanced machine learning
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline: formation of the ability to optimize, deploy and scale advanced models of machine learning of various types in practical laboratories. As a result of studying the discipline, the following abilities of doctoral students will be formed: - perform sample analysis and probabilistic modeling in conjunction with mathematical methods; - develop and implement optimization algorithms for these models; - create new solutions for machine learning; - implement and evaluate general machine learning models with reinforcement; - develop and implement recommendation systems. Within the discipline the following aspects will be considered: The study of linear separators. Probability and estimation, naive Bayes classifier. Generative and discriminatory classifiers. Logistic regression. Naive Bayes classifier. Kernel based approach. Generalization and retraining. Support vector machine. Busting Model selection. Linear regression. Active learning. Teaching with partial involvement of the teacher. Teaching without a teacher. Reduction dimension. Online training. Training with reinforcements.
-
Numerical methods for scientific computing tasks
- Type of control: [RK1+MT+RK2+Exam] (100)
- Description: The purpose of discipline:formation of the ability to perform the construction of algorithms and analysis of methods for the numerical solution of nonlinear differential equations of practical interest. As a result of studying the discipline, the following abilities of doctoral students will be formed: - describe the fundamental approaches in the field of numerical analysis and scientific calculations; - develop new computational algorithms; - analyze methods to ensure accurate solutions in the shortest possible time; - effectively solve optimal control problems using partial differential equations; - solve scientific computational problems using parallel and highly efficient computations. Within the discipline the following aspects will be considered: Difference approximation of simple differential operators. Various methods for constructing finite-difference schemes. Convergence. Consistency Sustainability. Density of mass distribution in continuous medium. Navier – Stokes equations. Dimensionless Navier – Stokes equations. Methods for modeling turbulent flows. Reynolds number. Navier-Stokes equations in cylindrical coordinate systems. Navier-Stokes equations in spherical coordinate systems. Wave equation. Two-step Lax-Wendroff method. Heat equation Simple explicit method. Richardson Method. Simple implicit method. Method Krenk - Nicholson. Burgers equation. Splitting method by physical parameters. Numerical methods for calculating processes in the boundary layer.
-
Pedagogical
- Type of control: Защита практики
- Description: Aim оf discipline: formation of the ability to carry out educational activities in universities, to design the educational process and conduct certain types of training sessions using innovative educational technologies.
-
Research
- Type of control: Защита практики
- Description: The purpose of the practice: gaining experience in the study of an actual scientific problem, expand the professional knowledge gained in the learning process, and developing practical skills for conducting independent scientific work. The practice is aimed at developing the skills of research, analysis and application of economic knowledge.