PhD in Computer Science

The course will facilitate the understanding of research planning and development, while enhancing the PhD student skills for reading, interpreting, writing and presenting his/her scientific and technical ideas to the peer community. The course will also deliver theory and examples of a variety of research methods and ethical principles associated to them.

Course learning outcomes
1. Identify the core responsible conduct of research topic areas.
2. Recognize the roles of professional ethics, integrity, responsibility and accountability.
3. Understand research profile.
4. Demonstrate effective scientific communication oration skills.
5. Understand how to conduct mentorship and collaboration with integrity.

This course develops doctoral students' ability to read, evaluate, organize, and synthesize current research literature in a critical and methodologically transparent way. By the end of the course, students will be able to formulate a focused review question, design and document an appropriate literature search strategy, critically compare published studies, identify gaps and contradictions in the field of computer science, and produce a concise survey paper.

Course learning outcomes
1. Define a focused review topic in computer science and explain its significance.
2. Critically compare research papers with respect to problem formulation, methodology, validation, assumptions, and limitations.
3. Write a concise survey paper that integrates and critically discusses recent literature in a subfield of computer science.
4. Critically assess peer work and provide constructive, professional feedback.

This course is designed to monitor progress and develop understandings, skills, and outlooks to conduct original, independent research at the PhD level. The student will develop (with the advisor's guidance) a research plan at the beginning of the semester that will state a research problem/question/hypothesis, its background, outline a research strategy and experimental approach, method of data collection, interpretation and validation, and method of communication of the project results to others. The research plan is used as the basis for assessment of the student's research progress.

Course learning outcomes
1. Discover and pursue a unique topic of research in order to construct new knowledge.
2. Design and conduct an original research project.
3. Develop skills in designing a discipline specific research methodology.
4. Develop a working knowledge of relevant literature in the thesis discipline.
5. Practice academic writing appropriate to the discipline and learn how to participate in the peer review process.
6. Discuss research and other topics with academics in the thesis discipline.

CSCI 701 is an accelerated course in the evolution and design of imperative programming languages, and modern software development techniques. In the first part of the course, we will use the C language to review the fundamentals of programming, and progress on to the concepts of data types, encapsulation, and reusability. Algorithms and performance will also be topics of discussion. While investigating the advantages and drawbacks of C, we will motivate some of the key features of object-oriented languages, such as inheritance and polymorphism, as well as generics. We will then look at the Java language as an instance of the OO paradigm, and discuss good OO design principles. In second part of the course, we will pull several of these threads together, where students will be asked to develop a mid to large-sized software project in the Java language.

Course learning outcomes
1. Explain why concepts such as abstraction and encapsulation are important in modern software design, and demonstrate with real examples.
2. Explain and show how component-based design and object-oriented approaches to software development can be used to enforce good design principles.
3. Perform asymptotic analysis and amortization to demonstrate the efficiency of an algorithm or data structure.
4. Perform rigorous functional and performance evaluations through testing.
5. Describe the key issues relating to the design and implementation of a new programming language.
6. Implement a simple functional or imperative programming language from a language specification.
7. Convert simple functional or imperative code into assembly.

This is an introductory PhD level course on the theory of computation and language theory. Since the students who have enrolled in our PhD program have various backgrounds, this course aims to give them a good foundation in theoretical computer science concepts and prepare them for upcoming courses. In the course, we will investigate different models of computation, such as finite-state machines and push-down automata, and discuss their strengths and limitations. We will then look at the historical and theoretical background of Turing machines and computability, and we will investigate ways to demonstrate (un-)decidability and (non-)recognizability of languages. Computational intractability and NP-completeness will also be discussed as part of the course.

Course learning outcomes
1. Formulate regular expressions, finite automata, context-free grammars, and pushdown automata to describe formal languages.
2. Analyze the expressive power and limitations of computational models and formal language classes.
3. Design Turing machines and explain their role as a universal model of computation.
4. Analyze computational problems and determine whether the corresponding languages are decidable, recognizable, or unrecognizable, providing rigorous justification.
5. Apply diagonalization, reductions, and related proof techniques to establish undecidability and non-recognizability results.
6. Analyze computational problems using major complexity classes, including P, NP, and PSPACE, and prove NP-completeness using polynomial-time reductions.
7. Construct and communicate rigorous mathematical proofs and arguments in computability and complexity theory.

In this course, students will be introduced to the most salient concepts in quantum computing, a relatively recent approach that specifically targets computing on the atomic and subatomic levels. The basics of quantum computing, including quantum states, unitary operators, and measurements will be presented. Through reversible and quantum circuits, students will learn and understand the advantages and limitations of quantum computing. Finally, different models of quantum computing will be given as illustrations of potential technological implementations.

Course learning outcomes
1. Derive residual quantum state after measurement, apply unitary operations on quantum states prove correctness of quantum algorithms.
2. Design quantum algorithms, describe quantum circuits and analyze the circuit complexity.
3. Explain quantum Fourier transform, construct quantum Fourier transform circuit, Understand phase estimation, describe the quantum algorithm solving phase estimation, explain order finding, describe and analyze the quantum order finding algorithm based on phase estimation, understand reflection and rotation, describe Grover's search algorithm and geometric analysis.
4. Define mixed states, describe teleportation and CHSH game, distinguish pure states vs. mixed states, understand bit-flip and phase-flip errors, construct simple quantum error correction codes, explain the need and principle of fault-tolerant quantum computing.

This course aims to provide Ph.D students with the necessary skills and knowledge for modern communication systems and computer networks with the main focus on wireless technologies. Practical network design, management, and control-related problems are considered together with wireless technologies such as cellular networks (including 3G, 4G, 5G, wireless LAN) and infrastructure-less technologies (Ad-Hoc Networks, WSNs). Distinct concepts and protocols in computer/communication network architectures such as addressing schemes, wireless variants of well-known routing and transport layer protocols, mobile IP and network design issues are considered. Case studies are provided for the application areas for wired/wireless technologies such as sensor networks, cloud computing, and the Internet of things (IoT).

Course learning outcomes
1. Review the distinct concepts and protocols of the wired/wireless network architectures and reflect on how they all fit together.
2. Assess the ways in which the technology of computer networks and wireless communication systems affect the suitability for particular applications.
3. Infer the implications of the specific modes of operation.
4. Test, simulate, and solve typical systems and their abstractions including analytical models for the quality of service analysis.
5. Review and assess security measures and countermeasures.
6. Adapt routing and IP addressing strategies in the design of computer networks to mobile systems.
7. Develop analysis and synthesis skills through a project relevant to the course.

Recent advances in technology have led to orders-of-magnitude increases in the amount of data that is being produced and archived. This phenomenon has been generically referred to as Big Data. Examples of such data include internet traffic and logs, social media, large scale scientific projects, and health records. The process of deriving actionable insights from these large volume and heterogeneous data sets is referred to as Big Data Analytics, and poses new challenges in all aspects of computing. This course will review essential statistical and machine learning techniques for data processing and examine their scalability. Students will gain an understanding of current best practices in Big Data processing and complete assignments and projects with a variety of relevant software tools.

Course learning outcomes
1. Have the required mathematical basis to study big data analytics.
2. Derive a whole range of covered algorithms from first principles.
3. Understand their theoretical basis and assumptions these may have.
4. Implement a range of techniques themselves in python from scratch.
5. Identify appropriate techniques for a given dataset.
6. Validate their approach, e.g. by means of cross-validation.
7. Understand the bias-variance trade-off (i.e. overfitting/underfitting).

This course aims to provide the students with a strong theoretical background and a large set of practical computational skills that are necessary for building up a competitive edge in utilizing modern trading strategies and technologies to capitalize on today's sophisticated financial markets and trading instruments. The lectures will provide a holistic coverage of financial market structure and dynamics, electronic trading infrastructure, software tools, and algorithmic trading strategies along with all necessary complementary terminology and information such as the common practices of both global and local markets. The course distinctively unites the domains of finance, econometrics and particularly computer science. Industry standards such as the Financial Information Exchange (FIX) protocol, limit-order book dynamics, algorithmic side of financial transactions, quantitative strategies, portfolio construction and management, news handling, and high-frequency trading methods will be covered along with practical examples that make heavy use of data mining, machine learning and time-series analysis techniques.

Course learning outcomes
1. Become comfortable and fluent with the terminology and practices of the financial applications of computer science and artificial intelligence.
2. Design and develop both client and server sides of trading applications using native Financial Information Exchange protocol.
3. Understand the strengths and weaknesses of major trading algorithms, and to assess and compare their performances.
4. Develop quantitative financial solutions which make use of statistics, artificial intelligence and computer science practices.
5. Be ready and well-prepared to participate, work in and/or scientifically contribute to financial organizations and industries.

This course aims at enabling PhD students to include cognitive aspects in the modeling of interactive systems in order to capture human errors that would be overlooked when verifying the system in isolation or in an ad hoc interaction context. The course starts with an introduction to cognitive science and a presentation of the human capabilities and limitations and how they affect the interaction process. It then illustrates the use of traditional cognitive models as well as new research approaches. Finally, cognitive modelling is integrated in a software engineering context in order to support the verification of interactive systems. Students' assessment includes practical modeling exercises, a term project applying the methods learned to a real-world problem, along with a presentation and examinations.

Course learning outcomes
1. Articulate the interdisciplinary aspects of HCI.
2. Link a user's cognitive capabilities and limitation to the user's task, determine the critical aspects of the task and assess their complexity.
3. Assess the impact of technology usage on the original user's task.
4. Predict the right level of abstraction at which to model a given interactive system.
5. Model systems and analyze system properties through the use of specialized tools.
6. Abstract, model and analyze real-world interactive systems in the context of a practical case study.
7. Develop writing and oral communication skills through a project relevant to the course.

Today's modern and digital world houses vast sources of data which are increasing exponentially as we shift to industry 4.0 practices and plan to have almost every electronic device connected by the use of Internet of Things (IoT) technologies. Analyzing data to identify patterns of interest and to make future projections is a very important feedback for improving processes and procedures. This course teaches modern data mining solutions and illustrates how they are used in decision support systems which are becoming increasingly ubiquitous in all business and industrial sectors. The topics to be covered include knowledge representation, input preparation (preprocessing), fundamental data mining techniques (like probabilistic modeling, linear modeling, association mining, etc.), minimum description length (MDL) principle, validation techniques, decision trees, association rules, instance-based learning models, data transformations, data projections, Bayesian methods, ensemble learning techniques (bagging, boosting, stacking, etc.), visualization, and warehousing within the supervised, unsupervised and semi-supervised learning frameworks.

Course learning outcomes
1. Understand a large set of supervised and unsupervised data mining algorithms along with their advantages and limitations.
2. Select an appropriate algorithm for a given problem and to quantitatively evaluate the fitness of a model generated by the selected algorithm.
3. Develop practical and trustable algorithmic solutions to be used by decision makers.
4. Be ready and well-prepared to participate, work in and/or scientifically contribute to related organizations and industries.

This course is a one-semester course intended for graduate students in Computer Science PhD program. It introduces the students to the concepts and existing models and techniques of deep learning. The main themes of the course are benefits, properties and challenges of deep learning; introduction to machine learning and optimization; challenges and common approaches; regularization; convolutional neural networks; deep recurrent networks and sequence learning; generative adversarial networks; autoencoders; advanced deep learning topics; applications. The students will learn how to design a deep learning architecture. The course presents several common deep learning techniques and exposes the students in learning how to use deep learning in a very efficient manner. Students will gain experience through assignments and projects, with an emphasis on proper deep learning practices with appropriate applications.

Course learning outcomes
1. Develop solutions using deep learning with Python programming language and environments to solve problems and perform specified tasks, using proper hyperparameters and optimization techniques.
2. Understand and apply concepts related to deep learning, such as logistics regression; gradient descent; regularization; Convolutional Neural Networks; Deep Recurrent Networks; Transformers; Generative Adversarial Networks; Autoencoders.
3. Critically determine how to select a deep learning approach for a domain and task.
4. Gain knowledge about state-of-the-art solutions using deep learning.
5. Understand the strengths and weaknesses of various deep learning approaches.

This course focuses on the use of formal methods in the modelling, specification and analysis of systems. Various kinds of systems from a number of application domains, such as human-computer interaction, software systems, distributed systems, systems biology and ecology, will be considered. Students will learn how to utilize formal notations and automated tools for simulation and analysis.

Course learning outcomes
1. Articulate the utility of formal methods, and how they relate to and complement other system validation techniques.
2. Apply formal modeling techniques to practical case studies.
3. Formally define system properties using appropriate mathematical notations.
4. Carry out proofs by hand and with the support of automated tools such as model checkers.
5. Critically compare different formal approaches, and be able to choose the most appropriate one for a given situation.
6. Develop writing and oral communication skills through a project relevant to the course.

Students will learn a number of important topics in the field of computational biomedicine, namely the acquisition and analysis of biomedical data. Students will study various topics within the biomedical domain, including primers of anatomy, physiology as well as the basis of various biomedical imaging techniques, such as electroencephalography (EEG), near-infrared spectroscopy (NIRS), functional Magnetic Resonance Imaging (fMRI), electrocardiogram (ECG), electromyogram (EMG), among others. Signal processing tools, such as filter theory, artifact rejection, as well as PCA, CSP and ICA will be covered. Additionally, we will cover how uni- as well as multi-variate features can be employed for decoding. In addition, we will cover a number of machine learning tools commonly used in the biomedical and neuroscientific community, such as RLDA, SVM, CNN, among others. Practical computing sessions will be carried out, such that the presented knowledge can be put into practice. Finally, students will also carry out a hands-on project, where all research-related steps will be covered, such as creating a hypothesis, experimental design with python, data acquisition and finally acquired data will be disseminated, interpreted and presented.

Course learning outcomes
1. Know some basic anatomical and physiological concepts.
2. Understand the basis of various biomedical imaging techniques, such as EEG, NIRS, fMRI, ECG, EMG, among others.
3. Understand which different imaging correlates can be measured and analyzed by various imaging modalities.
4. Understand and implement a range of data analytical techniques, that are common to biomedical related data analysis in particular, but also for computational approached in biomedicine in general.
5. Design and conduct a small biomedical study, i.e. experimental design, implementation, conduction, data analysis, report writing.

This course is a one-semester course intended for PhD students in the Computer Science program. It enhances the student's knowledge in computer vision concepts and approaches such as image formation, camera imaging geometry, feature detection and matching, stereo, image classification, scene understanding, and deep learning. In class, we will develop the intuitions and mathematics of the vision approaches. We will expose graduate students to a number of real-world applications and innovative research that are important to our daily lives. To further differentiate between theory and practice, we will guide students through a series of research projects such that they will have the chance to implement cutting-edge computer vision algorithms. Students will also gain experience through research papers, with an emphasis on proper computer vision practices with appropriate applications.

Course learning outcomes
1. Understand and explain concepts related to computer vision such as image formation, feature detection and matching, stereo, image classification, scene understanding, and deep learning.
2. Recognize and describe both the theoretical and practical aspects of computing with images.
3. Successfully implement several interesting and cutting-edge vision algorithms which are widely used in real-world applications such as image retrieval, image captioning, face recognition, scene categorization, and activity recognition.
4. Understand the strengths and weaknesses of various computer vision approaches and apply them to new scenarios.
5. Read and understand scholarly articles in computer vision and come up with new ideas to solve the same problem, which may lead students to write their research paper in computer vision.

This course covers the latest advances in probabilistic, statistical and network-based computational models for real world problems such as computer vision and decision making. Topical coverage will begin at hidden Markov Models, through Markov networks, up to conditional random fields. Additionally, we will cover novel advances in artificial intelligence such as causal reasoning, learning from data to reason, building models by learning from data and model optimization.

Course learning outcomes
1. Utilize common probabilistic graphical models and analyze real-world problems, data.
2. Utilize efficiently the causal reasoning methods available to find problems in statistical reasoning.
3. Construct models from data, verify them and use them on test cases or other real-world situations.

This course is a one-semester course intended for Doctoral students in Computer Science. Students will gain a deep understanding of pivotal topics, including the intricacies and innovations of recent artificial intelligence architectures. The course will cover diverse generative models such as encoder-decoder architectures, contrastive learning, seq2seq models, attention/transformer mechanisms, diffusion models, and a variety of specialized loss functions. This comprehensive overview will equip students with a broad understanding of the key approaches in generative modeling. Students will engage in practical, hands-on projects utilizing real-world generative models to tackle contemporary challenges and applications, including computer vision, speech and natural language processing, graph mining, reinforcement learning, and trustworthy machine learning. In this course, students will be introduced to extensive, large-scale datasets pivotal to the field of Generative Artificial Intelligence. All practical components will be delivered using the Python programming language, leveraging the power and versatility of leading libraries, including TensorFlow, PyTorch, and Scikit-learn.

Course learning outcomes
1. Demonstrate a thorough understanding of the foundational principles and advanced concepts underpinning generative models and their associated architectures.
2. Apply knowledge of deep learning architectures, such as encoder-decoder structures, diffusion models, seq2seq, attention mechanisms, transformers, BERT, and GPT.
3. Critically assess the performance of various generative models, utilizing both quantitative metrics and qualitative insights, to determine their suitability for different applications.
4. Integrate current research findings, methodologies, and advancements into generative model development and problem-solving.
5. Exhibit the ability to independently explore, understand, and integrate new tools, techniques, and research developments in generative AI, demonstrating a commitment to lifelong learning.
6. Develop writing and oral communication skills through a project relevant to the course.

The primary aim of this doctoral-level course is to help students gain a solid understanding of Software Defined Networks. The course will focus on the decoupling of the control plane and data plane with high level programming abstractions and orchestration. Moreover, it will look at how the network operators can become more flexible in their management of complex network functions, including routing, traffic engineering, security, and resource control. In addition, students will critically evaluate current research trends in Software Defined Networks. Finally, it will provide hands-on experience and practice to students, preparing them for real-world Software Defined Network implementations using Mininet, Mininet_WiFi, and/or Containernet.

Course learning outcomes
1. Describe the key benefits of the separation of the control and data planes in Software Defined Networks (SDNs).
2. Explore an SDN data plane, control plane, and application plane.
3. Configure an SDN-friendly network emulator.
4. Analyze and evaluate the pros and cons of applying SDNs in WANs.
5. Critically assess current research trends and emerging technologies in the field of SDNs.
6. Apply and evaluate security challenges in the SDN paradigm.
7. Develop writing and oral communication skills through a project relevant to the course.

This is a one-semester course intended for doctoral students in Computer Science. This course explores the interdisciplinary field of intelligent systems, focusing on the design, implementation, and application of systems that demonstrate human-like intelligence. The course covers foundational theories, algorithms, and methodologies used in developing intelligent systems, along with practical applications across various domains. Through a combination of lectures, discussions, case studies, and hands-on projects, students will gain a comprehensive understanding of intelligent systems and the skills needed to design and deploy them effectively.

Course learning outcomes
1. Summarize the fundamental concepts and theories underlying intelligent systems.
2. Explore various approaches and algorithms used in artificial intelligence and machine learning.
3. Design intelligent programs capable of solving subtypes of the most typical of artificial intelligence.
4. Analyze and evaluate the performance of intelligent systems in different application domains.
5. Critically assess current research trends and emerging technologies in the field of intelligent systems.
6. Develop writing and oral communication skills through a project relevant to the course.