Become an expert in artificial intelligence. Experience the leadership of top scientists from renowned foreign universities. Choose from interesting topics supported by the industry. Study and research with all advantages of full-time employment at KInIT, join the team of exceptional researchers and experience a unique KInIT culture.

Four reasons why PhD at KInIT

1. Excel

You’ll become a top expert in your selected field of intelligent technologies. You’ll make a contribution in the field that matters. With attention on ethical responsibility.

2. Grow

When working on your dissertation, you’ll have access to a supervising team including industrial partners and/or a world expert from a respected research institution.

3. Connect

You’ll be teamed up with the best researchers in the region. You’ll gain invaluable experience and contacts from abroad that will boost your career growth either in industry or academy.

4. Enjoy

During your PhD study, you’ll be a full-time KInIT employee with all the benefits. You’ll experience a unique KInIT culture based on trust, openness and respect.

KInIT PhD program is run in partnership with

Deadlines and process

  1. Express your interest — by April 20
    If you are interested in PhD study at KInIT, let us know as soon as possible.
  2. Discover KInIT, find a topic/supervisor match — by May 20
    We’ll get back to you and organize a matchmaking for a research topic and/or supervisor.
  3. Submit the application — by May 28
    If you have found a potential match, apply for a study (at FIT VUT Brno).
    We’ll guide you with submitting the application.
  4. Wait for the decision. — by July 20
    Wait until the admission process finishes. We’ll let you know our final decision.
  5. Start at KInIT. If we choose you, we provide you with further information on the work contract at KInIT.
    PhD study at FIT VUT officially starts in September 1st, 2021.

How the PhD study at KInIT works

1

Work at KInIT & study at FIT VUT

Viac info Menej info

We provide PhD study in collaboration with the Faculty of Information Technology at VUT, Brno (Czechia). You’ll be enrolled at the faculty as an external doctoral student. After you’re accepted for study, you’ll work as a full-time employee at our institute. Your doctoral research will be your primary job role.

2

Carry out research in a selected topic

Viac info Menej info

You can choose from multiple interesting topics of your dissertation. They are based on different mentoring schemes:

  1. Research topics supported by industrial research partners
  2. Research topics mentored by top scientists from abroad
  3. Research topics connected to international projects

You’ll have access to an expert supervising team. Depending on the selected scheme, it can include industrial partners and/or a scientist from a renowned foreign university.

3

Join our research teams

Viac info Menej info

While working at our institute you’ll be a member of a KInIT research team. The teams collaborate on various projects or on other activities:

4

Change the world

Viac info Menej info

The rest is up to you! Turn your curiosity and passion for intelligent technologies into new discoveries or inventions making the world a better place.

Dissertation topics

Advanced topics in machine learning

Supervising team: Mária Bieliková (KInIT guarantor), Jana Kosecka (George Mason University), Branislav Kveton (Google Research), Peter Richtárik (KAUST), Martin Takáč (Lehigh University), Peter Tino (University of Birmingham)
Key words: machine learning, deep learning, learning theory, optimization, trustworthiness

Viac info Menej info

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), Advances in Neural Information Processing Systems (NeurIPS), International Joint Conference on AI (IJCAI), Conference on Learning Theory (COLT).

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, time series analysis, unsupervised learning)
  • Deep Learning (e.g., architectures, generative models, deep reinforcement learning)
  • Learning Theory (e.g., bandits, game theory, statistical learning theory)
  • Optimization (e.g., convex and non-convex optimization, matrix/tensor methods, sparsity)
  • Trustworthy Machine Learning (e.g., accountability, causality, fairness, privacy, robustness)

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

Supervising team
Mária Bieliková
Expert Researcher
Jana Kosecka
Professor, George Mason University, USA
Branislav Kveton
Machine learning scientist, Google Research, USA
Peter Richtárik
Professor, King Abdullah University of Science and Technology, Saudi Arabia
Martin Takáč
Associate Professor, Lehigh University, USA
Peter Tino
Professor, University of Birmingham, UK

Natural language processing

Supervising team: Marián Šimko (supervisor, KInIT), Jana Kosecka (George Mason University)
Key words: deep learning, multilingual learning, transparency, interpretability, information extraction, text classification

Viac info Menej info

Continuous increase of data available world-wide emphasizes the need of its automatic processing and understanding. Particular challenges are posed by the heterogeneous and unstructured nature of text content provided in natural language. Natural language processing (NLP) ranks among the most prospective subfields of artificial intelligence with great potential of innovative applications affecting everyday life.

Recent advances in neural networks and machine learning allowed to push efficiency and scope of natural language understanding and generation forward. Yet, there remain many research challenges related to particular subtasks, application domains and languages. Further research and various resulting phenomena exploration is necessary. Special attention is drawn by the issues of interpretability and transparency of NLP models or by novel paradigms of learning addressing the problem of low-resource languages.

Particularly interesting challenges include, but are not limited to:

  • Transfer/multilingual learning
  • Interpretability and transparency for NLP
  • Domain-specific information extraction, text classification
  • Low-resource language processing
  • Visual grounding of natural language, image captioning

Relevant publications:

Supervising team
Marián Šimko
Expert Researcher
Jana Kosecka
Professor, George Mason University, USA

Machine learning in security

Supervising team: Daniela Chudá (supervisor, KInIT), Eset
Key words: machine learning, security, malware detection, detection of anomalies,  dimensionality reduction

Viac info Menej info

The constant creation and collection of large amounts of data is also associated with security problems. There is an increasing need to address security-related problems via methods based on artificial intelligence and machine learning.

Research problems in the domain of information security include the analysis of security logs, detection of a wide range of anomalies in network communication or detection of malicious behavior. These problems are related to the tasks of dimensionality reduction, clustering and classification. Another challenge is the interpretability of machine learning in the domain of security and trust.

Relevant publications:

Supervising team
Daniela Chudá
Expert Researcher
Eset

Intelligent data analysis

Supervising team: Viera Rozinajová (supervisor, KInIT), Sféra, Softec
Key words: data analytics, machine learning, optimization, smart grid, energy sharing

Viac info Menej info

Due to deployment of new technologies in our daily lives, huge amounts of data of various types are generated continuously. These datasets need to be processed efficiently – the information contained in them often supports the correctness and accuracy of decision-making processes. The development of data analysis methods is therefore an important part of IT research. For various reasons, traditional processing methods are not generally applicable, so new approaches need to be sought. They are mostly based on artificial intelligence and machine learning.

The current research is focused on optimization problems in energy domain. Here the renewable energy sources, batteries and electromobiles change the classical one-way centralised electric grid into two-way distributed network. This fact raises a number of research questions and problems to solve, e.g. microgrid optimization. New ways of energy sharing among prosumers need to be proposed, but there are some other research challenges that need to be addressed in this area.

However, the subject of interest can be data analysis in a broader context – considering tasks of prediction, clustering, classification or detection of anomalies in different domains.

 Relevant publications:

Supervising team
Viera Rozinajová
Expert Researcher
Sféra
Softec

Metaheuristic optimization

Supervising team: Anna Bou Ezzeddine (supervisor, KInIT), Sféra, Softec
Key words: nature-inspired computing, optimization, metaheuristics, dynamic predictions, microgrid

Viac info Menej info

Intelligent meta-heuristic algorithms provide a convenient and practical solution to complex optimization problems.Within metaheuristics, nature-inspired algorithms are gradually coming to the forefront because they are intelligent, can learn and adapt to the environment, much like biological organisms. These algorithms are interesting to the scientific community mainly because of the increasing complexity of the problems that need to be solved. They allow processing tasks with incomplete, probabilistic information when making decisions. The consideration of uncertainty in the modelling process, reflecting uncertainty in the real-world problem, may impact model outputs and therefore the optimum objective function value. Thus uncertainty adds an additional feature into any global optimization search.

Most real-world optimization tasks are highly non-linear and multimodal, with various complex constraints. Finding an optimal solution, or even a suboptimal solution, is generally not an easy task. There are several application domains where research problems of optimization can be addressed. One of these domains is energy. Optimal dynamic predictions of energy production and consumption will enable real-time planning and management of microgrid operations.

Relevant publications:

Supervising team
Anna Bou Ezzeddine
Expert Researcher
Sféra
Softec

Recommender and adaptive web-based systems

Supervising team: Michal Kompan (supervisor, KInIT), Mária Bieliková (supervisor, KInIT), Peter Brusilovsky (University of Pittsburgh), Branislav Kveton (Google Research), Luigi’s Box
Key words: recommender systems, information retrieval, user modeling, personalization

Viac info Menej info

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:

  • Diversity, novelty, and serendipity of recommendations
  • Explaining recommendations
  • Fairness and justice in recommendations
  • Biases in the recommendations
  • Novel machine learning approaches for adaptive and recommender systems

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

Thesis in this topic will be advised either by an external mentor or by Luigi’s Box.

Relevant publications:

Supervising team
Michal Kompan
Expert Researcher
Mária Bieliková
Expert Researcher
Peter Brusilovsky
Professor, University of Pittsburgh, USA
Branislav Kveton
Machine learning scientist, Google Research, USA
Luigi’s Box

Machine learning with human in the loop

Supervising team: Jakub Šimko (supervisor, KInIT), Peter Brusilovsky (University of Pittsburgh), Jana Kosecka (George Mason University)
Key words: machine learning, crowdsourcing, human computation, data annotation, human-in-the-loop, collective intelligence

Viac info Menej info

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 modes (e.g. user feedback on automated translations). They are particularly useful for surgical improvements of training data through identification and resolving of border cases. This is also directly related to explainability and interpretability of models.

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), collective intelligence. Each of these fields (or combination thereof) presents opportunities for new discoveries. They border on computer science disciplines such as data visualization, user experience (usability in particular) and software engineering.

The application domains of machine learning with 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), support of manual/automated fact-checking and more.

Relevant publications:

Supervising team
Jakub Šimko
Expert Researcher
Peter Brusilovsky
Professor, University of Pittsburgh, USA
Jana Kosecka
Professor, George Mason University, USA

Addressing small labeled data in machine learning training

Supervising team: Mária Bieliková (supervisor, KInIT), Ivan Srba (KInIT)
Key words: machine learning, meta-learning, transfer-learning, weak supervision, semi-supervised learning, explainability, transparency

Viac info Menej info

Ubiquity of computing is nowadays generating vast datasets potentially usable for training machine learned models. Yet, these data mostly come not annotated, i.e. without labels necessary for their use for supervised (most conventional) training methods (e.g. classification). Conventional acquisition of proper labels using human force is a costly endeavour. The result is that proper labels are either unavailable, few or of poor quality.

To circumvent the issue, many approaches are emerging and are presently researched by many researchers: meta-learning, transfer-learning, weak-supervision, zero/one-shot learning, semi-supervised learning. Each of these fields (or combination thereof) presents opportunities for new discoveries. Orthogonal to this, explainability and interpretability of models is an important factor to consider and advances in this regard are welcome (generally in AI and particularly in the mentioned approaches).

There are several domains of applications, where research of methods and models for addressing small labeled data can be applied. These include (but are not limited to) false information detection and support of manual/automated fact-checking.

Regarding the domain of applications, in KInIT, we are striving to acquire multiple research grant funding projects related to them – KInIT is already a part of the consortium which aims to become a Central European Digital Media Observatory as a part of European initiative EDMO (European Digital Media Observatory).

Relevant publications:

Supervising team
Mária Bieliková
Expert Researcher
Ivan Srba
Senior Researcher

KInIT supervisors

Mária Bieliková
Expert Researcher
Anna Bou Ezzeddine
Expert Researcher
Daniela Chudá
Expert Researcher
Michal Kompan
Expert Researcher
Jakub Šimko
Expert Researcher
Marián Šimko
Expert Researcher
Viera Rozinajová
Expert Researcher

External mentors

Peter Brusilovsky is a Professor at the School of Computing and Information, University of Pittsburgh, where he directs the Personalized Adaptive Web Systems (PAWS) lab. His research is focused on user-centered intelligent systems in the areas of adaptive learning, recommender systems, and personalized health. He is a recipient of Alexander von Humboldt Fellowship, NSF CAREER Award, and Fulbright-Nokia Distinguished Chair. Peter served as the Editor-in-Chief of IEEE  Trans. on Learning Technologies, and a program chair for several conferences including RecSys.

Jana Kosecka is a Professor at the George Mason University. She is interested in computational models of vision systems, acquisition of static and dynamic models of environments by means of visual sensing, high-level semantic scene understanding and human-computer interaction. She held visiting positions at UC Berkeley, Stanford University, Google and Nokia Research, and served as Program chair, Area chair or senior member of editorial board for  leading conferences in the field CVPR, ICCV, ICRA.

Branislav Kveton is a machine learning scientist at Google Research. He proposes, analyzes, and applies algorithms that learn incrementally, run in real time, and converge to near-optimal solutions as they learn. He made several fundamental contributions to the field of multi-armed bandits. His earlier work focused on structured bandit problems with graphs, submodularity, and low-rank matrices, and ranked lists. His recent work focuses on making bandit algorithms practical

Peter Richtárik is a Professor of Computer Science & Mathematics at KAUST. He is one of the founders and a Fellow of the Alan Turing Institute. Through his work on randomized and distributed optimization algorithms, he has contributed to the foundations of machine learning and federated learning. He serves as an Area Chair of leading machine learning conferences, including NeurIPS, ICML and ICLR.

Martin Takáč is an Associate Professor in the Department of Industrial and Systems Engineering at Lehigh University, where he is a core faculty at the OptML Group. His current research interests include the design and analysis of algorithms for machine learning, applications of ML, optimization, HPC. He serves as an Area Chair of ML conferences such as AISTATS, ICML and NeurIPS.

Peter Tino is a Professor at the School of Computer Science, University of Birmingham, UK. He is interested in the interplay between mathematical modelling and machine learning (dynamical systems, probabilistic modelling, statistical pattern recognition, natural computation). Peter is interested in both foundational aspects and applications in interdisciplinary contexts (e.g. astrophysics, biomedical sciences, cognitive neuroscience).

He is a Fellow of the Alan Turing Institute and has served on editorial boards of leading journals such as IEEE TNNLS, IEEE TCYB, Neural Networks and Scientific Reports.

Industrial partners

For more than 30 years, ESET® has been developing industry-leading IT security solutions that protect businesses and consumers worldwide.

Our R&D activities focus on all aspects of cybersecurity – detection of threats, research of malware, APTs, exploits, phishing, spam, monitoring of botnets and targeted attacks, protection of mobile platforms, IoT environment and network infrastructure as well as zero-trust security. We do a lot of data science and machine learning on very complex and extensive datasets.

Research Blog: https://www.welivesecurity.com/

Luigi’s Box is a Bratislava-based company, providing an intuitive site-search and product discovery SaaS to its customers, primarily in e-commerce. With two of the co-founders having PhD degrees in Software Engineering, R&D is at the core of Luigi’s Box DNA. We strive to bring our clients smart, high-quality tools which are personalized, context-aware, and scalable.

SFÉRA, a.s. is a Slovak supplier of IT solutions with over 30 years of experience. Energy, transport and industry sectors are in the focus of company’s activities. SFÉRA achieved the most prominent position in the energy sector, where the company has proposed a comprehensive portfolio of IT solutions for all market participants. SFÉRA takes an active part in the professional energy community, field of research and development.

Road transportation and environmental quality are two big research areas at Softec. We have designed our own IoT stations for winter road maintenance and air/noise pollution monitoring. The PhD study should focus on meaningful interpretation and enrichment of the measured data, as well as optimisation and prescriptive analytics with high added value, such as routing of smart road maintenance vehicles, spatial extrapolation of road condition forecasts and air pollution forecasting.

We are ready to provide data from our network of IoT stations measuring meteorological and environmental quality and conditions such as air and road temperatures, air humidity and pressure, air pollutant concentrations, noise level, etc.