Academically oriented

Combined study

4 years

prof. Ing. Lukáš Sekanina, Ph.D.

Chairman :
prof. Ing. Lukáš Sekanina, Ph.D.
Councillor internal :
doc. Ing. Jan Kořenek, Ph.D.
prof. Ing. Pavel Václavek, Ph.D.
prof. Ing. Tomáš Vojnar, Ph.D.
prof. Dr. Ing. Jan Černocký
prof. RNDr. Milan Češka, CSc.
prof. Ing. Adam Herout, Ph.D.
prof. Ing. Jan M. Honzík, CSc.
prof. Ing. Tomáš Hruška, CSc.
prof. Dr. Ing. Pavel Zemčík, dr. h. c.
prof. RNDr. Alexandr Meduna, CSc.
prof. Dr. Ing. Zbyněk Raida
prof. RNDr. Josef Šlapal, CSc.
Councillor external :
prof. Ing. Jiří Sochor, CSc.
prof.,RNDr. Jiří Barnat, Ph.D.

The goal of the doctoral degree programme is to provide outstanding graduates from the master degree programme with a specialised university education of the highest level in certain fields of computer science and information technology, including especially the areas of information systems, computer-based systems and computer networks, computer graphics and multimedia, and intelligent systems. The education obtained within this degree programme also comprises a training and attestation for scientific work.

• Graduates from the doctoral study programme are trained to independently work in research, development, or management.
• They are able to solve and/or to lead teams solving advanced conceptual, research, development, or production problems in the area of contemporary information technology and its applications.
• They can be engaged to work on creative tasks, to lead research and development teams, or to work in management of companies or organizations whenever there are required abilities to work in an independent and creative way, to analyze complex problems, and to propose and realize new and original solutions. Graduates from the doctoral study programme can also teach and/or scientifically work at universities.

FIT graduates in general and FIT doctoral graduates in particular do not have a problem finding employment at scientific, pedagogical or management positions both in Czech Republic and abroad.

• FIT   graduates of the doctoral study are capable of independent scientific, research and management work in the field of Informatics, Computer Technology and Information Technologies. Graduates are ready to solve challenging conceptual, research and development problems. They can independently conduct research, development and production in the field of modern information technology.
• Typically, they work as creative workers at top scientific research workplaces, as leaders of research and development teams and in scientific and pedagogical work at universities. Graduates of this program are also employed in higher functional positions of larger institutions and companies, where the ability to work independently, analyze complex problems and design and implement new, original solutions is required.
• And, last but not least, graduates typically continue as so-called "postdoc" in their academic careers in Czech Republic or abroad.

The requirements that the doctoral students have to fulfil are given by their individual study plans, which specify the courses that they have to complete, their presupposed study visits and active participation at scientific conferences, and their minimum pedagogical activities within the bachelor and master degree programmes of the faculty. A successful completion of the doctoral studies is conditional on the following:

• The student has to pass a doctoral state examination within which he/she has to prove a deep knowledge of methodologies, theories, and their applications in accordance with the state of the art in the areas of science that are given by the courses included in his/her individual study plan and by the theme of his/her future dissertation thesis. The doctoral state examination also encompasses an evaluation of the presumed goals of the future dissertation thesis of the student, of the chosen solution method, and of the so far obtained original results.
• The student has further to prepare and defend his dissertation thesis.

The rules are determined by the directions of the dean for preparing the individual study plan of a doctoral student.  The plan is to be based on the theme of his/her future dissertation thesis and it is to be approved by the board of the branch.

• obligatory doctoral study programme Courses, the total number of courses a student has to complete and their mapping into particular semesters.
• a Research Plan Content (brief descrition of research content - focuse at the intended research area and the doctoral thesis topic
• a Research Plan ( list of research activities focused at the intended research area and the doctoral thesis topic - conferences and seminars to be attended , work to be published)
• teaching duty according to BUT study rules and regulations
• doctoral study schedule

https://www.fit.vut.cz/fit/info/smernice/sm2018-13-en.pdf

Brno university of technology provides studies for persons with health disabilities according to section 21 par. 1 e) of the Act no. 111/1998, about universities and about the change and supplementing other laws (Higher Education Act) as amended, and according to the requirements in this field arising from Government Regulation No. 274/2016 Coll., on standards for accreditation in higher education, provides services for study applicants and students with specific needs within the scope and in form corresponding with the specification stated in Annex III to Rules for allocation of a financial contribution and funding for public universities by the Ministry of Education, Youth and Sports, specifying financing additional costs of studies for students with specific needs.

Services for students with specific needs at BUT are carried out through the activities of specialized workplace - Alfons counselling center, which is a part of BUT Lifelong Learning Institute - Student counselling section.

Counselling center activities and rules for making studies accessible are guaranteed by the university through a valid Rector's directive 11/2017 concerning the status of study applicants and students with specific needs at BUT. This internal standard guarantees minimal stadards of provided services.
Services of the counselling center are offered to all study applicants and students with any and all types of health disabilities stated in the Methodological standard of the Ministry of Education, Youth and Sports.

The study programme builds on both the ongoing follow-up Master's programme in Information Technology and the new follow-up Master's programme in Information Technology and Artificial Intelligence.
Students can also, according to their needs and outside their formalized studies, take courses and trainings related to the methodology of scientific work, publishing and citation skills, ethics, pedagogy and soft skills organized by BUT or other institutions.

1. Addressing Limitations of Large Language Models

   Large language models (LLMs) are powerful tools that can support a wide range of downstream tasks. They can be used e.g. in advanced conversational interfaces or in various tasks that involve retrieval, classification, generation, and more. Such tasks can be approached through zero-shot or few-shot in-context learning, or by fine-tuning the LLM on larger datasets (typically using parameter-efficient techniques to reduce memory and storage requirements). Despite their unprecedented performance in many tasks, LLMs suffer from several significant limitations that currently hinder their safe and widespread use in many domains. These limitations include tendencies to generate responses not supported by the training corpus or input context (hallucination), difficulties in handling extremely long contexts (e.g., entire books), and limited ability to utilize other data modalities such as vision, where state-of-the-art models generally struggle to recognize fine-grained concepts. The goal of this research is to explore such limitations, and – after selecting one or two of them to focus on – to propose new strategies to mitigate them. These strategies may include e.g.: • Shifting the generation mode closer to retrieval-style approaches and non-parametric language models; • Augmenting models with self-correction mechanisms and self-evaluation pipelines; • Efficiently supporting extended contexts; • Fuller utilization of multimodality, especially in the context of vision-language models; explainability analysis of models and the design of new training mechanisms supporting the ability to recognize fine-grained visual concepts as well; • Introducing novel fine-tuning techniques; • Improving and further utilizing the reasoning abilities of LLMs. Relevant publications: • Srba, I., Pecher, B., Tomlein, M., Moro, R., Stefancova, E., Simko, J. and Bielikova, M., 2022, July. Monant medical misinformation dataset: Mapping articles to fact-checked claims. In Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval (pp. 2949-2959). https://dl.acm.org/doi/10.1145/3477495.3531726 • Pikuliak, M., Srba, I., Moro, R., Hromadka, T., Smolen, T., Melisek, M., Vykopal, I., Simko, J., Podrouzek, J. and Bielikova, M., 2023. Multilingual Previously Fact-Checked Claim Retrieval. https://arxiv.org/abs/2305.07991 The application domain can be for example support for fact-checking and disinformation combatting, where the factuality of LLM outputs is absolutely critical. The research will be performed at the Kempelen Institute of Intelligent Technologies (KInIT) in Bratislava in cooperation with researchers from highly respected research units. A combined (external) form of study and employment at KInIT is expected.

   Tutor: Gregor Michal, doc. Ing., Ph.D.

1. round (applications submitted from 26.02.2024 to 31.05.2024)

1. Advanced algorithms of Video, Image, and Signal processing

   The topic concerns algorithms of image, video, and/or signal processing. Its main goal is to research and in-depth analyze existing algorithms and discover new ones so that they have desirable features and so that they are possible to efficiently implement. Such efficient implementation can be but does not necessarily have to be part of the work but it is important to prepare the algorithms so that they can be efficiently implemented e.g. in CPU, in CPU with acceleration through SSE instructions, in embeded systems, even in combination with FPGA, in Intel Xeon PHI, in extremely low power systems, or in other environments. It is possible to exploit algorithms of artificial intelligence, such as neural networks, especially CNNs The application possibilities of the algorithms are also important and the application can be but does not have to be part of the work. The algorithms/applications of interest include:

   • recognition of scene contents, events, and general semantics of video sequences (such as identification of traffic situations, identification in scenes in moview, action identification, etc.),
   • classification of video sequences using machine learning (AI)through deep convolution networks neural network or similar approaches (e.g. for industrial quality inspection, object of scene characteristics search, etc.), possibly in combination with object tracking in video using modern methods, 
   • parallel analysis of video and signal (e.g. for detection of coincidence of occurrence of object in video and characteristic signal shape in surveillance applications), fusion of video and sognals,
   • modern algorithms of video, image, and/or signal exploiting "client/server" or "cloud" approaches suitable e.g. for mobile technology and/or embedded systems,
   • algorithms of video compression and analysis through frequency or wavelet transforms or similar methods...

   After mutual agreement, individually selected algorithms can be considered as well as soon as they do belong to the general topic.

   Collaboration on grant projects, such as TACR, MPO, H2020, ECSEL (possible employment or scholarship).

   Tutor: Zemčík Pavel, prof. Dr. Ing., dr. h. c.

2. Advanced Methods of Computational Photography

   The project is concerned with advanced methods of computational photography. The aim is to research new computational photography methods, which comprises software solutions potentially supported by new optics and/or hardware. Our interest is on HDR image and video processing, color-to-grayscale conversions, spectral imaging, generative AI and others.

   • Further information: http://cadik.posvete.cz/tmo/
   • Contact: http://cadik.posvete.cz/
   • Cooperation and research visits with leading research labs are possible (Adobe Research, USA, MPII Saarbrücken, Germany, Disney Research Zurich, Switzerland, INRIA Bordeaux, France)

   Tutor: Čadík Martin, doc. Ing., Ph.D.

3. Advanced Rendering Methods

   The project is concerned with advanced rendering and global illumination methods. The aim is to research new photorealistic (physically accurate) as well as non-photorealistic (NPR) simulations of interaction of light with the 3D scene. Cooperation and research visits with leading research labs are possible (Adobe, USA, MPII Saarbrücken, Německo, Disney Curych, Švýcarsko, INRIA Bordeaux, Francie).
   
   

   • More information: http://cadik.posvete.cz/

   Tutor: Čadík Martin, doc. Ing., Ph.D.

4. Advanced topics in machine learning

   Machine learning is in the centre of research of artificial intelligence. Many researchers worldwide are dealing with the topics related to machine learning, both in academia and industry. This very dynamic field is characterized with fast transfer of solutions into practical use.

   The topics in this domain are defined by premier scientific conferences, where top-class researchers meet, for example ICML (International Conference on Machine Learning), NeurIPS (Advances in Neural Information Processing Systems), IJCAI (International Joint Conference on AI), COLT (Conference on Learning Theory).

   This thesis will be advised by an external mentor, who will also define its particular topic.

   Interesting research challenges are contained within (but are not limited to) these topics:

   • General Machine Learning (e.g., active learning, clustering, online learning, ranking, reinforcement learning, semi-supervised learning, unsupervised learning)
   • Deep Learning (e.g., architectures, generative models, deep reinforcement learning)
   • Learning Theory (e.g., bandits, statistical learning theory)
   • Optimization (e.g., matrix/tensor methods, sparsity)
   • Trustworthy Machine Learning (e.g., fairness, robustness)

   There are many application domains, where advanced machine learning methods can be deployed.

   The research will be performed at the Kempelen Institute of Intelligent Technologies (KInIT, https://kinit.sk) in Bratislava in cooperation with researchers from highly respected research units. A combined (external) form of study and full employment at KInIT is expected.

   Tutor: Bieliková Mária, prof. Ing., Ph.D.

5. Analysis of attacks on wireless networks

   The dissertation focuses on the security of wireless local area networks. The student should become familiar with selected wireless networks and their security as part of the solution. This work aims to study the theory of wireless networks, their properties, and possibilities of attacks, test the basic types of attacks, design new protection methods, conduct experiments, evaluate the results, and propose the direction of further research.

   Participation in relevant international conferences and publication in scientific journals are expected.

   
   Co-supervised by Dr. Kamil Malinka.

   Tutor: Hanáček Petr, doc. Dr. Ing.

6. Assessment of mental stress, anxiety and depression from analysis of brain signals

   Problem Statement: Mental stress, anxiety and depression are  mental health conditions that often occur together. In such a case, the person is stressed and is not able to control the worry, and it correspondingly affects his/her social and occupational activities. Hence, proper assessment and diagnosis for mental stress, anxiety and depression is required in order for a person to effectively keep taking part in his/her normal daily tasks and activities.

   Issues with Current Solutions: Unfortunately, conventional assessment and diagnostic measures are subjective in nature and are used only when the symptoms are already evident due to advanced stages of mental stress, anxiety and depression. However, mental stress, anxiety and depression do not occur overnight, rather it is a long process. Hence, detection of symptoms is required at early stages of mental stress, anxiety and depression because that may result in a cure or at least it will delay the onset of serious mental health issues associated with them.

   Challenges: Unlike other diseases where the symptoms like fever and cough allow people to seek help, symptoms at early stages of mental stress and anxiety are not easily identifiable. Hence, the brain needs to be continuously monitored for any sign of change or deterioration in order to detect the symptoms at early stage.

   Solution: The solution lies in the development of an objective and quantitative method that can detect mental stress, anxiety and depression at an early stage. Perception of mental stress, anxiety and depression originates in the brain; therefore, this research investigates the neurophysiological features extracted from brain electroencephalogram (EEG) signal to measure mental stress, anxiety and depression at early stage. This will require development of method for extraction of features as well as pattern recognition approach to provide a solution. The EEG dataset is already available for this project.

   Few Words About Supervision: I have extensive experience of working in the field of neuro-signal and neuroimage processing and I am currently head of a research group in this area. This is a multidisciplinary project and it will involve working with clinicians. However, the core of the project is related to IT in terms of development of a new method. Please feel free to contact me at malik@fit.vutbr.cz

   Tutor: Malik Aamir Saeed, doc., Ph.D.

7. Computer-aided Creativity

   Cílem disertační práce je výzkum v oblasti tzv. generativní umělé inteligence - ať už se jedná o difusní a adversiální modely pro generování videa, generativní textové modely pro vytváření příběhů, automatické generování počítačového kódu, hudby, reprezentaci znalostí fyziky a chemie a podporu vědecké kreativity, případně kombinace všech těchto přístupů. Práce se zaměří na řešení problémů interakce člověka s generovanými mezivýsledky, přirozeného označování jednotlivých částí a konceptů tak, aby bylo možné na průběžné výsledky navazovat, a na vývoj metod úpravy datových sad a postupů učení, aby bylo možné řešit společenské problémy, spojené s vytvářenými kreativními modely - otázky spravedlivosti modelů, předpojatosti a začlenění konceptů tzv. zodpovědné umělé inteligence.

   Tutor: Smrž Pavel, doc. RNDr., Ph.D.

8. Conversational Agents Combining Structural Knowledge and Learning from Text

   Konverzační agenti se pomalu stávají běžnou součástí rozhraní pro (prvotní) komunikaci se zákazníkem a odpovídání na jeho otázky. Výzkum v oblasti počítačového zpracování přirozeného jazyka se zaměřuje na vytvoření automatické klasifikace prvotní komunikace a, zejména, otázky uživatele, do předem daných tříd, k nimž existují konkrétní texty. Není však uspokojivě vyřešeno rozšiřování "znalostí" komunikačních agentů při aktualizaci strukturovaných dat, případně při přidání dalších textových materiálů.
   
   Cílem disertační práce je rozvinout existující přístupy využívající obrovské kolekce neanotovaných textových dat a způsoby kombinování strukturované a nestrukturované znalosti a optimalizace procesů při rozšiřování funkcionality stávajících i nových konverzačních agentů. Součástí práce bude i aplikace zkoumaných metod v rámci evropských projektů, na jejichž řešení se školitel podílí.

   Tutor: Smrž Pavel, doc. RNDr., Ph.D.

9. Cybersecurity aspects of the Internet of Things

   The dissertation focuses on the security of IoT systems. This work aims to study the theory of IoT systems, their properties, and possibilities of attacks, test the basic types of attacks, design a new method of protection, conduct experiments, evaluate results, and design further research.

   Participation in relevant international conferences and publication in scientific journals are expected.

   Co-supervised by Dr. Kamil Malinka.

   Tutor: Hanáček Petr, doc. Dr. Ing.

10. Development of neuromarker(s) for assessment of alcohol addiction

    Problem Statement: Alcohol addiction is a chronic and complex brain disorder causing devastating individual and social problems. Additionally, alcohol causes 3.3 million deaths a year worldwide, close to 6% of all deaths. Many of these deaths are associated with alcohol addiction. Therefore, it's important to look into methods for the diagnosis as well as the treatment of alcohol addiction.

    Issues with Current Solutions: Conventionally, screening and assessment of alcohol-related problems are mainly based on self-test reports. However, the accuracy of self-test reports has been questioned, especially for heavy drinkers, because the self-test reports may misguide the diagnosis due to the patient's memory loss (the patients cannot measure their alcohol consumption) and/ or dishonest behavior. Therefore, this research proposes to develop an objective and quantitative method for the detection of alcohol addiction.

    Challenges: As alcohol addiction results in changes in brain dynamics, hence, it is vital to investigate and develop a method based on brain activity. However, the main challenge in developing such an objective and quantitative method lies in its implementation for screening in smaller clinical setups. This limits the investigation to electroencephalogram (EEG) which is low cost, highly mobile and has good temporal resolution. Other modalities like MRI, PET etc are not feasible to be employed in smaller clinical settings.

    Solution: With current innovations in brain EEG signals, the brain pathways involved in addiction can be investigated. In the last few decades, EEG research has been used to understand the complex underlying processes associated with the pathophysiology of addiction. Interpreting such processes using brain networks using EEG can not only help in diagnosing addiction but also assist in treating addiction. This research aims to develop neuromarker(s) based on brain network interpretation using EEG. The neuromarker will involve the features extraction and corresponding development of the machine learning model.

    Few Words About Supervision: I have recently moved to FIT at Brno University of Technology. I have decade long experience of working in the field of neuro-signal and neuroimage processing and I am currently in the process of setting up a research group in this area. This is a multidisciplinary project and it will involve working with clinicians. However, the core of the project is related to IT in terms of development of a new method. Please feel free to contact me at malik@fit.vutbr.cz

    Tutor: Malik Aamir Saeed, doc., Ph.D.

11. Embedded Systems for Video/Signal Processing

    The topic focuses embedded image, video and/or signal processing. Its main goal is to research capabilities of "smart" and "small" units that have such features that allow for their applications requiring smyll, hidden, distributed, low power, mechanically or climatically stressed systems suitable of processing of some signal input. Exploitation of such systems is perspective and wide and also client/server and/or cloud systems. The units themselves can be based on CPU/DSP/GPU, programmable hardware, or their combination. Smart cameras can be considered as well. Applications of interest include:

    • classification of images or objects using machine learning (AI) using traditional methods or through deep convolution networks neural network or similar approaches (e.g. for industrial quality inspection, etc.),
    • parallel analysis of signal(s) and video (e.g. for robust detection of occurrence of object in industrial or surveillance applications),
    • modern algorithms of video, image, and/or signal exploiting "client/server" or "cloud" (with focus on the technlogy) suitable e.g. for mobile technology and/or embedded systems,
    • other similar topics can be individually consulted and considered.

    A possibility exists in collaboration on grant projects, especially the newly submitted TAČR, H2020, ECSEL ones (potentially employment or scholarship possible).

    Tutor: Zemčík Pavel, prof. Dr. Ing., dr. h. c.

12. Energy-aware Embedded Inteligence

    Cílem disertační práce je výzkum modelů vestavěné inteligence, která explicitně pracuje s energetickou náročností konkrétních operací a optimalizuje svoji činnost na základě konkrétních omezení na straně jednotlivých zařízení, případně celého systému. Součástí bude i realizace vybraných modelů na vhodném typu hardware, který bude možné využít v mezinárodních projektech, na jejichž řešení se vedoucí podílí.

    Tutor: Smrž Pavel, doc. RNDr., Ph.D.

13. Evolutionary Optimization of Logic Circuits

    Ukazuje se, že metody syntézy číslicových obvodů využívající evolučních algoritmů, zejména kartézského genetického programování pracujícího přímo nad reprezentací na úrovni hradel, jsou schopny produkovat implementace, které jsou v řadě případů mnohem efektivnější (typicky kompaktnější) nežli implementace získané pomocí současných syntézních technik využívajících interní reprezentace (např. AIG) a iterativní aplikace deterministických přepisovacích pravidel. Typickým cílem optimalizace je redukovat počet hradel optimalizovaného obvodu. V praxi se však vyskytuje požadavek optimalizovat obvod z hlediska více kriterií (např. zpoždění, plocha na čipu). V případě využití systému pro účely resyntézy je multikriteriální optimalizace nutností z důvodu zachování zpoždění obvodu, jehož část je předmětem optimalizace. 
    
    Cílem disertační práce je navázat na předchozí výzkum a zabývat se možnostmi multikriteriální optimalizace číslicových obvodů s ohledem na dobrou škálovatelnost. Dále se předpokládá využití alternativních reprezentací jako je např. majority uzel, které lépe odrážejí principy nových technologií.
    
    Výzkum spadá do témat řešených výzkumnou skupinou Evolvable Hardware.

    Tutor: Vašíček Zdeněk, doc. Ing., Ph.D.

14. Explainable Artificial Intelligence

    Použití některých metod strojového učení, například v poslední době populárních hlubokých neuronových sítí, přináší problémy architektury tzv. černé skříňky, která sice může v některých případech správně rozhodovat, ale není možné snadno interpretovat způsob rozhodování, ověřovat, v jakém kontextu jsou závěry ještě věrohodné a nakolik mohou vést drobné změny vstupu ke zcela jiným závěrům.
    
    Cílem disertační práce je rozvinout existující přístupy k měření "dokazatelně správných" modelů umělých neuronových sítí a propojit je s technikami generování konfliktních (adversarial) příkladů, aby bylo možné kontrolovat a revidovat existující řešení, využívaná v praxi. Součástí práce bude i aplikace zkoumaných metod v rámci evropských projektů, na jejichž řešení se školitel podílí.

    Tutor: Smrž Pavel, doc. RNDr., Ph.D.

15. Generative AI and Image Processing using Neural Networks

    The project is concerned with advanced methods of image processing and generative AI. The aim is to research new methods using machine learning, in particular deep neural networks.

    • Contact: http://cadik.posvete.cz/
    • Cooperation and research visits with leading research labs are possible (Adobe Research, USA, MPII Saarbrücken, Germany, Disney Research Zurich, Switzerland, INRIA Bordeaux, France)

    Tutor: Čadík Martin, doc. Ing., Ph.D.

16. Human-AI collaboration in dataset creation

    The models created in machine learning can only be as good as the data on which they are trained. Researchers and practitioners thus strive to provide their training processes with the best data possible. It is not uncommon to spend much human effort in achieving upfront good general data quality (e.g. through annotation). Yet sometimes, upfront dataset preparation cannot be done properly, sufficiently or at all.

    In such cases the solutions, colloquially denoted as human-in-the-loop solutions, employ the human effort in improving the machine learned models through actions taken during the training process and/or during the deployment of the models (e.g. user feedback on automated translations). They are particularly useful for surgical improvements of training data through identification and resolving of border cases.

    Human-in-the-loop approaches draw from a wide palette of techniques, including active and interactive learning, human computation, and crowdsourcing (also with motivation schemes of gamification and serious games). With recent emergence of large language models (LLM), the original human-in-the-loop techniques can be further boosted to create extensive synthetic training sets with comparatively small human effort.

    The domains of application of human-in-the-loop are predominantly those with a lot of heterogeneity and volatility of data. Such domains include online false information detection, online information spreading (including spreading of narratives or memes), auditing of social media algorithms and their tendencies for disinformation spreading, support of manual/automated fact-checking and more.

    Relevant publications:

    • Cegin, J., Simko, J. and Brusilovsky, P., 2023. ChatGPT to Replace Crowdsourcing of Paraphrases for Intent Classification: Higher Diversity and Comparable Model Robustness. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing https://arxiv.org/pdf/2305.12947.pdf
    • J. Šimko and M. Bieliková. Semantic Acquisition Games: Harnessing Manpower for Creating Semantics. 1st Edition. Springer Int. Publ. Switzerland. 150 p. https://link.springer.com/book/10.1007/978-3-319-06115-3

    The research will be performed at the Kempelen Institute of Intelligent Technologies (KInIT, https://kinit.sk) in Bratislava in cooperation with industrial partners or researchers from highly respected research units. A combined (external) form of study and full employment at KInIT is expected.

    Tutor: Šimko Jakub, doc. Ing., PhD.

17. Image and video quality assessment

    The project deals with image and video quality assessment metrics (IQM). The aim is to explore new ways how to incorporate human visual system properties into IQM. In particular, we will consider perception of HDR images, and utilization of additional knowledge (in form of metadata, 3D information, etc.) about the tested scenes using machine learning (e.g. neural networks).

    • Further information: http://cadik.posvete.cz/iqm/
    • Contact: http://cadik.posvete.cz/
    • Cooperation and research visits with leading research labs are possible (MPII Saarbrücken, Germany, Disney Research Zurich, Switzerland)

    Tutor: Čadík Martin, doc. Ing., Ph.D.

18. Improving natural language processing

    The recent development of large language models (LLMs) shows the potential of deep learning and artificial neural networks for many natural language processing (NLP) tasks. Advances in their automation have a significant impact on a plethora of innovative applications affecting everyday life.

    Although large-scale language models have been successfully used to solve a large number of tasks, several research challenges remain. These may be related with individual natural language processing tasks, application domains, or the languages themselves. In addition, new challenges stemming from the nature of large language models and the so-called black-box nature of neural network-based models.

    Further research and exploration of related phenomena is needed, with special attention to the problem of trustworthiness in NLP or new learning paradigms addressing the problem of low availability of resources needed for learning (low-resource NLP).

    Interesting research challenges that can be addressed within the topic include:

    • Large language models and their properties (e.g., hallucination understanding)
    • Trustworthy NLP (e.g., bias mitigation, explainability of models)
    • Adapting large language models to a specific context and task (e.g. via PEFT, RAG)
    • Advanced learning techniques (e.g., transfer learning, multilingual learning)
    • Domain-specific information extraction and text classification (e.g., novel methods for sentiment analysis, improving conversation quality in chatbots)

    Relevant publications:

    • Pikuliak, M., et al. SlovakBERT: Slovak Masked Language Model. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 7156–7168, ACL, 2022
      http://dx.doi.org/10.18653/v1/2022.findings-emnlp.530
    • Pikuliak, M., Šimko, M. Average Is Not Enough: Caveats of Multilingual Evaluation. In Proceedings of the The 2nd Workshop on Multi-lingual Representation Learning (MRL), pages 125–133, ACL, 2022
      http://dx.doi.org/10.18653/v1/2022.mrl-1.13

    The research will be performed at the Kempelen Institute of Intelligent Technologies (KInIT, https://kinit.sk) in Bratislava in cooperation with industrial partners or researchers from highly respected research units from abroad. A combined (external) form of study and full employment at KInIT is expected.

    Tutor: Šimko Marián, doc. Ing., Ph.D.

19. Improving performance of Large Language Models for downstream tasks

    Large language models (LLMs) are increasingly being used for a wide range of downstream tasks where they often show a good performance in zero/few-shot settings compared to specialized fine-tuned models, especially for tasks in which the LLMs can tap into the vast knowledge learned by them during the pre-training. However, they lag behind the specialized fine-tuned models in tasks requiring a more specific domain knowledge and adaptation. Additionally, they often suffer from problems such as hallucinations, i.e., outputting coherent, but factually false or nonsensical answers; or generating text laden with biases propagated from pre-training data. Various approaches have recently been proposed to address these issues, such as improved prompting strategies including in-context learning, retrieval-augmented generation or adapting the LLMs through efficient fine-tuning.

    Each of these approaches (or combination thereof) presents opportunities for new discoveries. Orthogonal to this, there are multiple important factors of models like their level of alignment with human values, their robustness, explainability or interpretability and advances in this regard are welcome as well (generally in AI and particularly in the mentioned approaches).

    There are many downstream tasks, where research of the LLM adaptation methods can be applied. These include (but are not limited to) false information (disinformation) detection, credibility signals detection, auditing of social media algorithms and their tendencies for disinformation spreading, and support of manual/automated fact-checking.

    Relevant publications:

    • Macko, D., Moro, R., Uchendu, A., Lucas, J.S., Yamashita, M., Pikuliak, M., Srba, I., Le, T., Lee, D., Simko, J. and Bielikova, M., 2023. MULTITuDE: Large-Scale Multilingual Machine-Generated Text Detection Benchmark. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing https://arxiv.org/abs/2310.13606
    • Vykopal, I., Pikuliak, M., Srba, I., Moro, R., Macko, D., and Bielikova, M., 2023. Disinformation Capabilities of Large Language Models. Preprint at arXiv: https://arxiv.org/abs/2311.08838

    
    The research will be performed at the Kempelen Institute of Intelligent Technologies (KInIT, https://kinit.sk) in Bratislava in collaboration with industrial partners or researchers from highly respected research units involved in international projects. A combined (external) form of study and full employment at KInIT is expected.

    Tutor: Bieliková Mária, prof. Ing., Ph.D.

20. Information Extraction from the WWW

    The topic of identifying and extracting specific information from documents on the Web has been the subject of intensive research for quite a long time. The basic obstacles that make this problem difficult are the loose structure of HTML documents and absence of meta-information (annotations) useful for recognizing the content semantics. This missing information is therefore compensated by the analysis of various aspects of web documents that include especially the following:

    • Document HTML code (DOM)
    • Document Text (Keyword Search, Statistical Text Analysis, Natural Language Processing Methods)
    • Visual organization (page content layout, visual properties)

    A background knowledge about the target domain and the commonly used presentation patterns is also necessary for successful information extraction. This knowledge allows a more precise recognition of the individual information fields in the document body.

    Current approaches to information extraction from web documents focus mainly modeling and analyzing the documents themselves; modeling the target information for more precise recognition has not yet been examined in detail in this context. The assumed goals of the dissertation are therefore the following:

    • Analysis of existing domain models such as UML class diagrams, E-R diagrams or ontology.
    • Extending these models with the specification of recognizing particular data in documents (e.g. regular expressions, advanced text classification).
    • Design of information extraction methods based on a comparison of the structure of the information presented in the document and the expected structure of the target information.

    Experimental implementation of the proposed methods using existing tools and experimental evaluation on real-world documents available on the WWW is also an integral part of the solution.

    Tutor: Burget Radek, doc. Ing., Ph.D.

21. Machine Learning for Information Identification on the Web

    Although there are technologies that allow publishing data on the WWW in machine-readable form (such as JSON-LD, RDFa, etc.), a large amount of structured data is still published on the web in the form of plain HTML/CSS code, which greatly limits the possibilities of their further use.
    
    Recently, new machine learning methods (especially deep learning methods) are gaining importance, which show interesting results, e.g., in recognizing important entities in weakly structured or unstructured data (e.g., text or images). However, the area of web document processing has not received much attention from this perspective. Existing works deal with the identification of simple data items and neglect structured data and more complex usage scenarios.
    
    The goal of this topic is to analyze and develop web content models suitable as input for machine learning and, at the same time, machine learning methods suitable for recognizing structured data in web documents.

    Tutor: Burget Radek, doc. Ing., Ph.D.

22. Machine Learning Models that Reason

    V poslední době se stává čím dál patrnější, že k překlenutí propasti mezi současnými nejlepšími modely strojového učení a lidským učením nestačí jen zvyšovat počty parametrů a čekat na výkonnější hardware, který zvládne zpracování bilionů parametrů. Zdá se, že je třeba hledat nové modely, schopné objevovat a uvažovat na úrovni vysokoúrovňových pojmů a vztahů mezi nimi.
    
    
    
    Cílem disertační práce je výzkum nových modelů strojového učení, které překonají potřebu enormního množství příkladů, které jsou potřeba pro naučení chování, zvládnutelného lidmi velmi rychle (například počítač potřebuje sehrát velké množství her ke zvládnutí jednoduché videohry, zatímco člověk to zvládne velmi rychle, lidé ze sady proměnných rychle určí, jaká je příčinná souvislost mezi nimi, dokáží argumentovat sledem úvah atd.), a omezí problém sebejistého chybování (overconfident incorrectness) současných modelů. Budou zkoumány postupy učení, přidávající iterativně nové relevantní informace a také metody, podporující přímé pravděpodobnostní odvozování. Výsledky budou demonstrovány na vybraných problémech, zahrnujících mj. vysvětlování videa či tvorbu inferenčních grafů, operujících nad pojmy a vztahy mezi nimi.

    Tutor: Smrž Pavel, doc. RNDr., Ph.D.

23. Measuring output quality of large language models

    The advent of large language models (LLMs) is raising research questions about how to measure quality and properties of their outputs. Such measures are needed for benchmarking, model improvements or prompt engineering. Some evaluation techniques pertain to specific domains and scenarios of use (e.g., how accurate are the answers to factual questions in such and such domain? how well can we use the generated answers to train a model for a specific task?), others are more general (e.g., what is the diversity of paraphrases generated by an LLM? how easy to detect it is that the content is generated?).

    Through replication studies, benchmarking experiments, metric design, prompt engineering and other approaches, the candidate will advance the methods and experimental methodologies of LLM output quality measurement. Of particular interest are two general scenarios:

    1. Dataset generation and/or augmentation, where LLMs are prompted with (comparatively small) sets of seeds to create much larger datasets. Such an approach can be very useful, when dealing with a domain/task with limited availability of original (labelled) training data (such as disinformation detection).
    2. Detection of generated content, where stylometric-based, deep learning-based, statistics-based, or hybrid methods are used to estimate whether a piece of content was generated or modified by a machine. The detection ability is crucial for many real-world scenarios (e.g., detection of disinformation or frauds), but feeds back also to research methodologies (e.g., detecting the presence of generated content in published datasets or in crowdsourced data).

    The candidate will select (but will not be limited to) one of the two general scenarios, identify, and refine specific research questions and experimentally answer them.

    Relevant publications:

    • Cegin, J., Simko, J. and Brusilovsky, P., 2023. ChatGPT to Replace Crowdsourcing of Paraphrases for Intent Classification: Higher Diversity and Comparable Model Robustness. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing https://arxiv.org/pdf/2305.12947.pdf
    • Macko, D., Moro, R., Uchendu, A., Lucas, J.S., Yamashita, M., Pikuliak, M., Srba, I., Le, T., Lee, D., Simko, J. and Bielikova, M., 2023. MULTITuDE: Large-Scale Multilingual Machine-Generated Text Detection Benchmark. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing https://arxiv.org/pdf/2310.13606.pdf

    The research will be performed at the Kempelen Institute of Intelligent Technologies (KInIT, https://kinit.sk) in Bratislava in cooperation with industrial partners or researchers from highly respected research units. A combined (external) form of study and full employment at KInIT is expected.

    Tutor: Šimko Jakub, doc. Ing., PhD.

24. Modern Algorithms of Computer Graphics

    The topic concerns algorithms of computer graphics and image synthesis. Its main goal is to research new algorithms so that their features and application possibilities are better understood so that they are improved or newly created.  If suitable, it is possible to work on various platforms, includeing parallel CPUs, such as x86/64, ARM, Xeon PHI, GPU, etc. or other cores in CUDA, OpenCl, VHDL, etc. Algorithms of interest include:

    • rendering using selected computer graphics methods (such as ray tracing, photon mapping, direct rendering of "point clouds", etc.),
    • modeling of scenes and redering using artificial intelligence, including image synthesis using neural netowrks (especially CNNs),
    • processing and rendering of "lightfield" images, their acquisition, or possibly compression, reconstruction of 3D scenes from images and/or video, eventually also fusing with other sensors, such as LIDAR or RADAR,
    • modern algorithms of geometry suitable for applications in cpmputer graphics and perhaps also 3D printing,
    • emerging algorithms of 3D synthesis, holography, wavelet transform, frequency transform, etc.

    After mutual agreement, individually selected algorithms can be considered as well as soon as they do belong to the general topic.

    Collaboration on grant projects, such as TACR, H2020, ECSEL possible (employment or scholarship).

    Tutor: Zemčík Pavel, prof. Dr. Ing., dr. h. c.

25. Multimodal analysis for assessment of mental health

    Problem Statement: The importance of mental health has increased significantly over the past decade. However, the methods for the assessment of mental health issues at early stages are still in their infancy compared to the availability of corresponding methods for early assessment of physical health issues. Hence, it is required that due research is done to develop methods for early assessment of abnormalities leading to mental health problems.

    Issues with Current Solutions: Unlike physical health parameters, the mental health is assessed through a number of subjective parameters. Hence, there is lack of objective and quantitative methods for mental health assessments. In addition, the patients seek help when their mental health problem is at advanced stage. So, there is lack of continuous monitoring for mental health issues.

    Challenges: Many of the abnormalities related to mental health issues are subtle in nature and are related to behavior and other changes in facial expressions, speech and handwriting. In addition, there are changes in cortisol levels, skin conductance, heart rate variability and breathing rate. Hence, there are multiple modalities that should be included for measuring and quantifying any abnormalities related to mental health.

    Solution: Every modality has its pros and cons. For example, in neuroimaging, functional magnetic resonance imaging has high spatial resolution (in mm) and low temporal resolution (in seconds) while electroencephalogram has low spatial resolution (in cm) and high temporal resolution (in milliseconds). Combining both of them will result in high spatial as well as high temporal resolution. This research deals with the assessment of abnormalities leading to mental health problems by utilizing multimodal approach. The various modalities may include, but not limited to, electroencephalogram (EEG) brain signals, facial videos, speech audios, handwriting and text from social media. The physiological parameters from various modalities include, but not limited to, the heart rate, breathing rate, dominant emotion, fatigue and stress. Dominant emotion can be classified as positive or negative and then sub-classified as sad, happy, angry etc. Data mining and data fusion techniques will be developed for this multimodal analysis. The corresponding multimodal data is available for this project.

    Few Words About Supervision: I have extensive experience of working in the field of neuro-signal and neuroimage processing and I am currently head of a research group in this area. This is a multidisciplinary project and it will involve working with clinicians. However, the core of the project is related to IT in terms of development of a new method. Please feel free to contact me at malik@fit.vutbr.cz

    Tutor: Malik Aamir Saeed, doc., Ph.D.

26. Prediction of second stroke from data collected using standard stroke protocol

    Problem Statement: Stroke is a condition in which the supply of blood to the brain is restricted or stopped. When a stroke patient arrives at the hospital, a standard protocol is followed to provide the medical assistance to the patient as well as to assess the affect of the stroke on the brain. Generally, a second stroke may follow which can be devastating for the patient. Hence, it is critical to predict the next stroke and provide care that can avoid the next stroke or at least minimize its affects.

    Issues with Current Solutions: The standard protocol at the hospital involves blood and urine tests as well as neuroimaging using CT and MRI scans. These tests are used to assess the damage done by the first stroke. However, the prediction of the next stroke depends on the doctor's experience and is very subjective. In many cases, the patient who is sent home after treatment, suffers the second more devastating stroke at home which can result in permanent disability.

    Challenges: The standard protocol at the hospitals result in generation of lot of data, for example, hundreds of images from CT and MRI scans, hundreds of enzymes from blood and urine tests etc. The challenge is to collectively analyze all of this data and find correlations that can predict the second stroke for the patient.

    Solution: This research will develop objective prediction method for occurrence of the second stroke from the data collected at the hospital using the standard protocol for stroke assessment, treatment and management. The analysis and development will involve machine learning techniques that handle multimodality data involving images, signals, text, and numbers.

    Few Words About Supervision: I have recently moved to FIT at Brno University of Technology. I have decade long experience of working in the field of neuro-signal and neuroimage processing and I am currently in the process of setting up a research group in this area. This is a multidisciplinary project and it will involve working with clinicians. However, the core of the project is related to IT in terms of development of a new method. Please feel free to contact me at malik@fit.vutbr.cz

    Tutor: Malik Aamir Saeed, doc., Ph.D.

27. Pre-onset detection of Alzheimer's Disease (AD) by investigating brain dynamics

    Problem Statement: Among all the types of dementia, Alzheimer's disease (AD) is the most common form with 70 % of those affected by dementia having AD. As the prevalence of AD increases with age, the number of people living with AD is expected to rise over the next decades due to better quality of life that results in increase in age across many countries. All this has resulted in an increased focus on ensuring pre-onset detection of AD and the corresponding intervention, which can lead to slowing the progression of the disease by providing adequate diagnostics.

    Issues with Current Solutions: Preclinical AD happens 10 to 15 years before the onset of the disease resulting in changes in the brain without showing any actual symptoms of the disease like memory loss etc. Pre-onset means detecting AD in or before the preclinical stage. The existing state-of-the-art methods mainly focus on the detection of later stages of AD, and the detection of preclinical AD is still an open research problem. Hence, this research targets pre-onset detection of AD (that is, early detection of Preclinical AD) because that will have huge impact on the lives of people. This can lead to early intervention and may result in further slowing the progression of the disease.

    Challenges:  At the stage of preclinical AD, the related signs and symptoms are not clear, and hence people at this stage do not seek any help. Therefore, a method for pre-onset detection of AD should be part of the regular health screening process and hence should be available in small clinical setups.

    Solution: Method for detection of preclinical AD will involve investigating underlying brain mechanisms to monitor and track changes related to pre-onset detection of AD. Magnetic resonance imaging (MRI) will be used as a reference to investigate the brain dynamics however it cannot be used in practice due to its high-cost and specialized setup environment which limits its usage at the screening stage. Electroencephalogram (EEG) will be used in this research which is widely available, is low cost, has a good temporal resolution, and has high mobility. Therefore, this project aims to investigate the changes in underlying brain mechanisms using EEG to develop EEG-based neuromarker for pre-onset detection of AD. The neuromarker will involve the features extraction and corresponding development of the machine learning model.

    Few Words About Supervision: I have recently moved to FIT at Brno University of Technology. I have decade long experience of working in the field of neuro-signal and neuroimage processing and I am currently in the process of setting up a research group in this area. This is a multidisciplinary project and it will involve working with clinicians. However, the core of the project is related to IT in terms of development of a new method. Please feel free to contact me at malik@fit.vutbr.cz

    Tutor: Malik Aamir Saeed, doc., Ph.D.

28. Recommender and adaptive web-based systems

    The recommender systems are an integral part of almost every modern Web application. Personalized, or at least adaptive, services have become a standard that is expected by the users in almost every domain (e.g., news, chatbots, social media, or search).

    Obviously, personalization has a great impact on the everyday life of hundreds of million users across many domains and applications. This results in a major challenge - to propose methods that are not only accurate but also trustworthy and fair. Such a goal offers plenty of research opportunities in many directions:

    • Novel machine learning approaches for adaptive and recommender systems
    • Trustworthy recommendation methods for multi-objective and multi-stakeholder environments
    • Explaining recommendations
    • Fairness and justice in recommendations
    • Biases in the recommendations

    
    There are several application domains where these research problems can be addressed, e.g., search, e-commerce, news, and many others.
    
    Relevant publications:

    • V. Bogina, T. Kuflik, D. Jannach, M. Bielikova, M. Kompan, C. Trattner. Considering temporal aspects in recommender systems: a survey. User Modeling and User-Adapted Interaction, 1-39, 2022. https://doi.org/10.1007/s11257-022-09335-w 
    • I. Srba, R. Moro, M. Tomlein, B. Pecher, J. Simko, E. Stefancova, M. Kompan, A. Hrckova, J. Podrouzek, A. Gavornik, and M. Bielikova. Auditing YouTube’s Recommendation Algorithm for Misinformation Filter Bubbles. ACM Trans. Recomm. Syst. 1, 1, Article 6, March 2023. https://doi.org/10.1145/3568392 

    
    The research will be performed at the Kempelen Institute of Intelligent Technologies (KInIT, https://kinit.sk) in Bratislava in cooperation with industrial partners or researchers from highly respected research units from abroad. A combined (external) form of study and full employment at KInIT is expected. 

    Tutor: Kompan Michal, doc. Ing., PhD.

29. Security impacts of artificial intelligence on cybersecurity

    In recent years, there has been a huge increase in the quality of the output of neural networks generating synthetic content, which has gone hand in hand with a significant simplification of the use of AI-based tools and their increased availability. Thus, the growing trend in the use of artificial intelligence brings new challenges to the field of cybersecurity. The most prominent examples are the use of "deepfakes" to attack biometric systems or the use of deep learning techniques to detect cyber attacks. The goal of this work is to analyze new trends, approaches, real attacks, their characteristics, impacts, and potential applications in a selected area of cybersecurity. The work should then propose new AI-based protection methods based on the analysis and research on the state of security for the selected areas.

    Recommended areas of focus for the thesis:
    Impact of generative AI on code and application security
    The human factor in AI-based attacks - e.g., increasing the ability of humans to recognize these types of attacks
    Impacts of generative AI on the security of biometric authentication (voice, face, ...)

    Participation in relevant international conferences and publication in scientific journals are expected.

    Co-supervised by Dr. Kamil Malinka.

    Tutor: Hanáček Petr, doc. Dr. Ing.

30. Semantic Cartesian Genetic Programming

    Zavedením sémantických operátorů do genetického programování umožnilo významně zefektivnit jeden ze základní stavebních pilířů evolučních výpočetních technik a zredukovat množství generací potřebných k nalezení řešení. Cílem disertační práce je navázat na předchozí výzkum a zabývat se možnostmi zavedení sémantických operátorů do kartézského genetického programování. 
    
    Výzkum spadá do témat řešených výzkumnou skupinou Evolvable Hardware.

    Tutor: Vašíček Zdeněk, doc. Ing., Ph.D.

31. Transformation of Formal Systems for Languages

    Tutor: Meduna Alexandr, prof. RNDr., CSc.

32. Usable security and privacy

    The growing trend in IT technology is increasing demands on users, who must make more and more decisions regarding IT security. As part of the solution of the thesis, there should be an introduction to security techniques and their usability. The thesis goal is to improve the usability of the selected security techniques to be effective in practice by concerning human factors knowledge and user-centered design principles. Primary interest of the work will be in the user perception of emerging technologies such as AI or changing trends in single and multi-factor authentication. Participation in relevant international conferences and publication in professional or scientific journals is expected.

    Co-supervised by Dr. Kamil Malinka.

    Tutor: Hanáček Petr, doc. Dr. Ing.

33. Visual Geo-Localization and Augmented Reality

    The project deals with geo-localization in unknown environments using computer vision and computer graphics methods. The aim is to investigate and develop new image registration techniques (with geo-localized image database or 3D terrain model). The goal is an efficient implementation of proposed methods on mobile devices as well as search for additional applications in the area of image processing, computational photography, and augmented reality.

    • Further information: http://cadik.posvete.cz/locate/
    • Contact: http://cadik.posvete.cz/
    • Cooperation and research visits with leading research labs are possible (Adobe Research, USA, MPII Saarbrücken, Germany, Disney Research Zurich, Switzerland, INRIA Bordeaux, France)

    Tutor: Čadík Martin, doc. Ing., Ph.D.

34. Tutor: Hanáček Petr, doc. Dr. Ing.

35. Tutor: Hanáček Petr, doc. Dr. Ing.

36. Tutor: Meduna Alexandr, prof. RNDr., CSc.

37. Tutor: Meduna Alexandr, prof. RNDr., CSc.

38. Tutor: Meduna Alexandr, prof. RNDr., CSc.

39. Tutor: Meduna Alexandr, prof. RNDr., CSc.