Studies start in September 2023

Deadline for applications: Tuesday, 16th of May 2023 (2 p.m.)

Submit Application

Information from the Department Secretary:

Contact Form |Tel. 25002533 | Fax: 25002633

One (1) post in the following field: “Explainable Deep Learning via Self-Attention Methods”

Description: Deep learning-based artificial intelligence (AI) models currently represent the state of the art for making functional predictions in a vibrant and expansive field of research. A neural network method called deep learning imitates how human brain is wired. However, it is frequently unknown what assumptions underlie the predictions made by predictive models.

The proposed thesis addresses "explainable AI," referring to a set of procedures and techniques that make it possible for human users to understand and believe the output produced by machine learning algorithms. We are particularly interested in self-attention mechanisms, which are deep learning techniques that add extra focus to a particular component. The thesis will focus on one component within the network's architecture that's responsible for managing and quantifying the interdependent relationships within input elements, called self-attention.

Prerequisites: MSc in Data Science, working experience in Python.

Research Advisors: Sotiris Chatzis, Associate Professor, sotirios.chatzis@cut.ac.cy

One (1) post in the following field: “Credible Artificial Intelligence Technologies”

Description: The adoption of Artificial Intelligence (AI) and digital transformation in general, has recently become a mantra for public organisations & public administration, private sector, the sector of education and training and at all level of governance. Digital literacy is essential for life in a digitalized world while developing an understanding of the risks and opportunities of emerging technologies and mainly those of AI is of great importance. Essential element is also the understanding of AIs implications around justice and governance, which should be accompanied by the ethical aspects of AI and should include a gender mainstreaming mindset, as women, although accounted for 54% of all tertiary students in the EU in 2017, they are yet particularly underrepresented in the digital sectors. The goal of this project is to develop novel guidelines for the ethical and responsible development of AI tools, including Large Language Models. 

Research Advisors: Sotiris Chatzis, Associate Professor, sotirios.chatzis@cut.ac.cy

One (1) post in the following field: “Magnetic Resonance Imaging (MRI) Guided Focused Ultrasound System for Ablation in the Abdominal Area”

Description: Focused ultrasound is a modality that can be used to treat various diseases in the area of oncology using thermal protocols. The thermal effects of Focused ultrasound can be monitored with excellent contrast using Magnetic resonance imaging (MRI).

The offered position will concentrate on the design and evaluation of 4-D MR compatible robotic system. A major task is to design an agar-based brain phantom.Simulations will be performed in order to optimize the focused ultrasound therapeutic protocols.A transducer design dedicated for ablation in the abdominal area will be performed. The successful applicant is expected to extensively evaluate the system in the developed phantom in the laboratory setting and inside an MRI scanner.MRI sequences will be optimized in order to monitor the thermal effects of ultrasound.

Required Qualifications: Degree in Electrical Engineering, or Mechanical Engineering, or Physics or Medicine or Biomedical Engineering

Research Advisors: Christakis Damianou, Professor,  christakis.damianou@cut.ac.cy

One (1) post in the following topic: “Intelligent Fault Diagnosis for Distributed Systems and Networks”

The emergence of 5G, IoT, WSN and Industry 4.0 technologies has made possible the collection of large amounts of real-time data about a monitored environment. There is currently a need to develop fault tolerant methods and architectures for distributed systems and networks. The proposed research will focus on innovative fault diagnosis approaches that can learn characteristics or system dynamics of the monitored environment and adapt their behavior in order to handle missing or inconsistent data and achieve fault tolerant monitoring. 

Required Qualifications: Undergraduate (BSc) and postgraduate (MSc) degrees in Electrical Engineering or Computer Science or related field. Prior research experience or specialization in related topics will be considered an advantage.

Funding is available for full-time qualified applicants through involvement in aerOS https://aeros-project.eu/, a Horizon Europe Research Program.

Research Advisors: Michalis Michaelides, Assistant Professor, michalis.michaelides@cut.ac.cy

One (1) post in the following topic: “Optimizing Machine Learning Pipelines over Large-scale Distributed Data Streams”

Deploying machine learning (ML) applications over distributed stream processing engines (DSPEs) such as Apache Spark Streaming, Heron, and Flink is a complex procedure that requires extensive tuning along two dimensions. First, DSPEs have a vast array of system configuration parameters (such as degree of parallelism, memory buffer sizes, etc.) that need to be optimized to achieve the desired levels of latency and/or throughput. Second, each ML model has its own set of hyper-parameters that need to be tuned as they significantly affect the overall prediction accuracy of the trained model. These two forms of tuning have been studied extensively in the literature but only in isolation from each other, while recent studies have shown complex interactions between the two. The goal is to develop novel combined system and ML tuning approaches that will automate the deployment of ML applications over DSPEs in an optimized way.

Required Qualifications: Undergraduate (BSc) and postgraduate (MSc) degrees in Computer Science or related field. The ideal candidate should enjoy working on cutting-edge systems research problems and have good software development skills. Prior research experience or specialization in related topics will be considered an advantage.

There is possibility of funding through involvement in a Research Program for qualified applicants or as teaching assistants.

Research Advisors: Herodotos Herodotou, Assistant Professor, herodotos.herodotou@cut.ac.cy

One (1) post in the following topic: “Machine Learning-based Management of Hybrid Distributed Data Storage Systems”

Description: The continuous improvements in memory, storage devices (e.g., NVRAM, SSDs), and network technologies of commodity hardware have led to the introduction and proliferation of hybrid (or multi-tiered) distributed data storage systems. However, the complexity of data movement across the storage tiers and caches increases significantly, making it harder for applications to take advantage of the higher I/O performance offered by the system. The goal is to develop novel machine learning-based algorithms that will automate all the relevant decisions made by such systems, such as cache eviction, admission, prefetching, and tier migrations.

Required Qualifications: Undergraduate (BSc) and postgraduate (MSc) degrees in Computer Science or related field. The ideal candidate should enjoy working on cutting-edge systems research problems and have good software development skills. Prior research experience or specialization in related topics will be considered an advantage.

There is possibility of funding through involvement in a Research Program for qualified applicants or as teaching assistants.

Research Advisors: Herodotos Herodotou, Assistant Professor, herodotos.herodotou@cut.ac.cy

One (1) post in the following topic: “Picture Archiving and Communication System (PACS) and advanced medical imaging techniques”

PACS is a system used in medical imaging to store, retrieve, distribute, and present digital images. It is widely used in healthcare facilities such as hospitals, clinics, and imaging centers. The position in this area involves research into developing and improving PACS systems such as designing and implementing new algorithms and tools, developing new methods for integrating multiple imaging modalities, and exploring new applications of PACS in clinical practice. The potential research topics include developing new methods for handling large volumes of imaging data, improving image quality and resolution, developing methods for integrating multiple imaging modalities – including visible light modalities - and exploring new applications of PACS in clinical research and education.

Candidates are expected to have a strong background in a related field, such as computer science, biomedical engineering or medical physics, as well as experience in medical imaging and software and database development.

Required Qualifications: Degree in Computer Science or Biomedical Engineering or Medical Physics.

Research Advisors: Efthyvoulos Kyriacou, Assistant Professor, efthyvoulos.kyriacou@cut.ac.cy

One (1) post in the following topic: “Advanced Medical Image and Video Analysis”

The offered position will focus on the analysis of medical images and videos. The main area will be analysis of carotid plaque images. The task will be the evaluation of the risk of stroke from the analysis of the plaque. The candidate will have to work on the application of new techniques on like AM-FM image analysis on real time video analysis of the movement of the plaques and the forces during a cardiac cycle. In addition, artificial intelligence and deep learning techniques will be applied in order to create models for classification.

The candidate(s) must have good programming skills mainly in Python language.

Required Qualifications: Degree in Computer Science/Engineering Electrical Engineering, biomedical engineering or a relevant field.

Research Advisors: Efthyvoulos Kyriacou, Assistant Professor, efthyvoulos.kyriacou@cut.ac.cy

One (1) post in the following topic: “Illicit art trade: Monitor, Combat, Restitution and Management”

Introduction: It has been estimated that 140,000-700,000 cultural goods are transacted in Europe annually, with a total value of 64-318 M€. The illicit art trade ranks as the third-largest type of black-market trafficking, behind drugs and weapons. It is a multi-parametric issue operating parallel to the legal market, and it concerns artefacts that come through theft, illicit excavation, and the production of fakes or forgeries.

This doctoral study is devoted to the research and formation of innovative and specialized sustainable strategies and tools to combat the illicit trafficking of cultural goods and activate mechanisms to monitor, trace, safeguard, and restitute. Additionally, the research seeks to create a network of communication and cooperation with domestic and foreign competent authorities in the field of combating the looting and illicit trafficking of cultural property.

Proposed Methodology and Scope: The objective is to develop specialized and in-depth research centering on the authentication, tracing, and identification process of cultural heritage objects and their management thereafter.

The approach revolves around the development of a best-practice manual and a legal framework relevant to each stage of the process. A significant focus of the PhD is the 3D digitization of cultural heritage, coding of 3D cultural objects, and the making of 3D duplicates and digital twins of cultural objects. The study will be partly based on the EU Study VIGIE2020/654 “Study on quality in 3D digitization of tangible cultural heritage” and will be conducted in close and direct collaboration with an interdisciplinary team of experts, including distinguished academics and scientists from local and international institutions.

The ultimate purpose is to create a special “code” system for the holistic 2D and 3D documentation of tangible cultural objects, to create an authentication code system in order to combat the illicit art trade, to protect and preserve cultural heritage and relevant materials, and to promote ethical trade.

Subjects and Procedure: Subjects and procedure to this doctorate thesis include authenticity, traceability, and identification of cultural heritage objects, the sharing of databases and inventories, making of 3D duplicates and digital twins of cultural objects, cultural heritage law and cultural diplomacy, archiving, documentation and registration of cultural heritage objects, and relevant data and metadata collection, processing, and usage. Additional focal points involve acquisition policies, provenance protocol and technology, and arbitration/mediation and Alternative Dispute Resolution.

Research Advisors: Marinos Ioannides, Assistant Professor, marinos.ioannides@cut.ac.cy

One (1) post in the following topic: “An approach for the digital holistic documentation in Cultural heritage: The extent of the intangible heritage embedded within in a digital format as a part of a holistic processing of Cultural Heritage data.”

Keywords: Data Processing, Digital Heritage, Metadata, Holistic, Cultural Heritage, Intangible, Paradata

Digitising cultural heritage assets requires the collection of vast quantities of complex multimodal data, images and recordings, in order to identify and document features both tangible features, e.g. materials, physical complexities and geographical context. By processing the data collected from the entirety of a heritage asset, the aim is to identify and record the tangible features, in order to access the intangible aspects, thus revealing its story - from production to final use. However, we face the issue of how we can identify and interact with this intangible story. Furthermore, to fully digitise the story and memory of a heritage asset we need to consider the challenge of how we can ensure the maintenance of links and relationships present between the various data.   

The aim of this research is to effectively and innovatively utilise and combine the broad spectrum of acquisition techniques (technological and literary) and the various data produced from them, to generate a high quality and accurate holistic documentation of heritage assets. By doing this the recording goes beyond the complexity of the tangible objects structure in order to extrapolate the intangible memory and value embedded within. Particular consideration will be made to the factors that define what is tangible and what is intangible heritage.

A key aspect of this PhD is to delve into the background and history of the asset and in this case jewellery and goldsmith's articles, with the aim to identify when and why it was made, for whom and how/why it was used or reused within its historic context, allowing us to reconstruct and interpret the hidden intangible story within the tangible. In particular we need to take into account the aspects that sit on the borderline of tangible and intangible heritage, which include production techniques (such as, skill level, knowledge of the manufacturer and available tools), use wear and reuse.

By developing a workflow to process the data collected in order to extract hidden meaning and store knowledge, a contribution can be made to a global standard for holistic documentation. As an additional part of this process, a key element will be the production of a metadata schema, such as vocabularies for the classification of cultural heritage assets to be input into a complex knowledge management system.  

Research Aims and Questions: An important aspect of this research is the processing and handling of the data collected, in order to ensure an accurate, accessible and searchable record of the metadata. Some of the challenges that will be considered, include: 

• How can we define and identify meaning through the recording of tangible assets?
• In which ways can we extract meaning from within tangible assets?
• Which tools and methodologies can be applied to record this intangible heritage and subsequent storytelling?  
• How can we develop a pipeline for the processing of information and data, considering the following stages: Information and Data – Knowledge– Interpretation – Meaning – Story – Presentation – Enrichment – (Story).
• In what ways can we enrich information and multimedia data formats to reconstruct the fragmentary story woven into the lifecycle of the asset? 
• What terminology is critical for the definition of metadata in order to create a classification and contribute to an ontology that ensures an accurate and usable record of cultural heritage assets?
• How can we contribute to the production of a widely adopted standard for the recording of metadata and have an added value and impact on the wider academic and scientific community?   

Research Advisors: Marinos Ioannides, Assistant Professor, marinos.ioannides@cut.ac.cy

One (1) post in the following topic: “The holistic approach to digital documentation of Cultural Heritage: The case of the Kyrenia Ship”

The digitization of cultural heritage constitutes one of the best contemporary practices for understanding and maintain the memory of the past, for the creation of a detailed digital database, for the protection and the promotion of cultural heritage, as well as for the accessibility by a wider audience. The digitization of cultural heritage represents a complex process depending on different aspects, such as the size and the type of the material under study, the location etc., while it is also consisted of several stages, with the data acquisition being the initial one. A necessary aspect of data collection is obtaining data based on the provenance, the purpose, the age, the artistic aspect, the value, and the uniqueness of the items under study. For the purpose of the digitization it is necessary to collect available data provided by different types of analyses in order to create a digital three-dimensional (3D) model, which constitutes the best way for a visual presentation of cultural heritage elements.

The necessary process applied for data acquisition produces a large number of data, metadata and paradata that need to be stored, combined and processed for the creation of the three-dimensional (3D) model. Paradata constitute the main component for a comprehensive and holistic documentation in order to describe the properties, the presentation and the visualisation within cultural and historical contexts of the object/artefact under study. Since paradata describe the human processes by which the object/artefact under study was created, they need to be recognised and understood during the process of acquiring data as an integral part of heritage visualisation practices. The usage and understanding of paradata results in the holistic documentation of the cultural heritage. Through the ‘storytelling’ of the object/artefact under study, the thoughts before its creation, the processes by which it was created and the final result (metadata), it maintains its value.

More specifically, as far as maritime cultural heritage is concerned, digitization contributes to the recording and the management of the maritime archaeological material, as well as to its holistic documentation. Moreover, the promotion and the accessibility of the material constitute great advantages of the digitization.

Thus, within the framework of the digitization of maritime cultural heritage, this PhD research proposal suggests the digitization of the Kyrenia Ship.

The digitization of the Kyrenia Ship constitutes a really complex issue, as well as a huge challenge, as the Ship is exhibited at the Ancient Shipwreck Museum in the Kyrenia Castle. Kyrenia is located at the northern part of Cyprus, which is under the illegal Turkish occupation since 1974. Therefore, the digitization of the Ship becomes imperative, not only for the documentation and the promotion of the Ship, but also for its accessibility by a wider audience.

Having as a main subject the digitization of the Kyrenia Ship, the goal of this PhD will be the creation of a unique system for data documentation, collection, storage and use of the acquired information.

One of the main challenges of the certain PhD will be the creation of a proposal for the selection of digitization methods to acquire data for data, metadata and paradata, depending on the cultural heritage element under study. Moreover, the quantity and the kind of data needed to tell the ‘story’ of the element will be under question, as well as the way for creating data that are collected and that can be useful for different user groups and easily accessible for further use.

Research Advisors: Marinos Ioannides, Assistant Professor, marinos.ioannides@cut.ac.cy

One (1) post in the following field: “Optimizing the Complexity of Knowledge. A cross-domain and FAIR-compliant approach for semantic modelling in Digital Cultural Heritage”

In the last decade, the evolution of the Internet of Things (IoT), open science infrastructures, and web application frameworks have opened up brand new scenarios for data circulation, discovery, extraction, and visualization. This change has increased the interest in applying W3C's semantic web technologies to develop and connect metadata and Digital Heritage data to the Linked Open Data (LOD) web. The digital humanities make an extensive and very specific use of these tools, which are primarily intended for two-dimensional (2D) file formats. Three-dimensional (3D) formats have recently been introduced in this scenario and present many challenges in terms of data, metadata, attribution, circulation, and linking.

Regardless of format, the data, information, and knowledge associated with the digital twin of a cultural object are difficult to keep together, although this could be mitigated in the early stages of acquisition and processing. Data modelling techniques allow us to approach the heterogeneity of data sources that characterise what we can define as Big Data of Cultural Heritage. However, to enable a comprehensive knowledge representation, it must be based on Knowledge Organization Systems (KOS) such as models, schemas and ontologies. Furthermore, current data modelling activities need to be framed within the FAIR (Findable, Accessible, Interoperable, Reusable) principles, which are already included in the H2020 data management guidelines and are a key requirement for the modern life cycle of scientific data.

The objectives of this research activity are:

• To produce high-fidelity, high-cross-section formal representations of the movable and immovable assets of Digital Cultural Heritage, incorporating their associated knowledge as intangible cultural heritage (ICH), as well as other interpretive aspects - e.g. cognitive aspects.
• Applying FAIR principles to different semantic data modelling workflows and necessarily meeting the requirements on data quality.

This innovative approach will be based on the lifecycle of holistic documentation, considering the different data types, structures, and end-user needs resulting from the ERA Chair Mnemosyne project and the selected case studies. In particular, this PhD project will be developed in conjunction with the expected results from the data processing of the Mnemosyne project. The results of the described project will be further explored in the form of guidelines and modular semantic web tools.

Research Advisors: Marinos Ioannides, Assistant Professor, marinos.ioannides@cut.ac.cy

One (1) post in the following field: “From the inside out: Developing a holistic documentation of Cultural Heritage objects, to manage the multi-faceted complexity of tangible elements, to identify the intangible history embedded within and how this can be represented in digital form."

The digitization of cultural heritage requires data collection that involves a wide range of variety and multimodal methods of data, images, and recordings in order to identify and document characteristics such as composition, physical uses, and social and cultural context. The aim is to collect and evaluate the data from tangible characteristics to access the intangible aspects, to discover its history. However, we have to consider how we can identify and interact with this intangible history. An equally important challenge is also how we can ensure the preservation and the connection that exists between the various data.

A key aspect of this PhD research is the socio-cultural background of the cultural heritage, aiming to identify when and why it was created, for whom, and how or why it was used, allowing us to reconstruct and interpret the hidden history within it. In particular, we need to consider aspects bordering on texture and heritage, which include production technique (such as skill level, construction knowledge, and available tools), and wear and tear. To address this challenge, we need to develop a new methodology and guidelines for data processing, through filtering, severance, and embedded data integration, to enrich the collected data to develop a holistic policy documentation that integrates tangible, and intangible artifacts, as well as a wider knowledge.

Throughout this PhD research, within the framework of the ERA Chair H2020 project 'MNEMOSYNE', the aim is to contribute to the overall documentation process, using selected case studies, to efficiently and innovatively utilize the data acquired through a wide range of acquisition techniques, alongside tools such as video, audio or multispectral and thermal imaging, to enhance existing knowledge and datasets and fully document the cultural asset. To obtain that, seek to capture the complexity of the structure and pre-existing knowledge to analyze and understand the story embedded within. This research will also focus to manage and sort the data, metadata, and deliverables, which will lead to a knowledge management system. In addition, it will aim to contribute to a manual for the management and structure of collected metadata and deliverables to ensure high-quality, comprehensive, and accurate data processing.

Research objectives and questions: An important aspect of this research is the processing of the collected data to ensure an accurate, accessible, and open record of the metadata. Some of the challenges that will be addressed include:

• How can we identify and record intangible heritage from recording physical assets and in particular, how can we apply different tools and methodologies to extract interpretation from digital resources?
• How can we visualize the data in a way that the story can emerge?
• In what ways can we modify and enrich data to recreate the story that connects to the cultural asset?
• How critical is the term for defining and classifying metadata to create an accurate and usable record?
• Working on the previous four points, how can we contribute to the production of a prototype for recording metadata?
• How can this research provide an ontology/knowledge system to support the complexity of cultural assets to be considere..

Research Advisors: Marinos Ioannides, Assistant Professor, marinos.ioannides@cut.ac.cy