Courses at EWSCS 2026

• Luca Aceto (professor at Reykjavik University, Iceland and Gran Sasso Science Institute, Italy)
  The equational logic of concurrent processes: Results and proof techniques   toggle abstract   (course details)

  This course will survey some of the results on the equational axiomatisation of notions of equivalence over algebras of concurrent processes that have been obtained over the last 45 years. It will focus on equational axiomatisations of various notions of process equivalence over fragments of Milner's Calculus of Communicating Systems, presenting both positive and negative results on the existence of (finite) equational characterisations of the resulting algebras of processes. The course will also highlight proof techniques for showing such results, various notions of completeness one might want to consider over algebras and some of the main challenging open problems in this research field.

  The course will assume no previous familiarity with equational logic, process algebra and universal algebra.

  ---

• Matej Pavlović (researcher at Near One)
  Distributed consensus, state machine replication, and Byzantine fault tolerance   toggle abstract   (course details)

  Starting from the basics of distributed computing, we explain the most commonly used abstractions and algorithms. We show why consensus is such an important and fundamental problem in this field, why it is impossible to achieve the way we would most like it, and what we do about it in practice. We explore how distributed computer systems can be made to tolerate failures of some of their components, what happens if malicious actors actively try to disrupt those systems, and when and how we can still guarantee correct operation.

  No specific prior knowledge is required.

  ---

• Miklós Simonovits (research professor at Alfréd Rényi Institute of Mathematics, Hungary)
  Extremal graph theory and related areas   toggle abstract   (course details)

  In the middle of the previous century Combinatorics and Graph Theory consisted only from a few larger areas and many isolated, puzzle-looking problems and results. Next a strange two-direction interaction was built up between these areas and Theoretical Computer Science. Finding fast algorithms became more and more important and understanding several deeper parts of Graph Theory, or more generally, Discrete Mathematics became very important. Random, modified, and pseudo-random methods were introduced by Paul Erdős, and also random codes were used already by Shannon. The investigation of when and how can the randomness be eliminated was also very important in the related cases.

  The aim of my first three lectures is to introduce the audience into one of the fastest developing areas of Discrete Mathematics, Extremal Graph Theory, and some further related areas. A sketch of my lectures is the following.

  1. Ramsey and Turán theorems
  2. Classification of Extremal problems, degenerate and non-degenerate extremal problems [6]
  3. Random graphs, the three models: hypergeometric, binomial models and the stopping rule model.
  4. Szemerédi Regularity Lemma, and its applications, see e.g. [5]
  5. Removal Lemma
  6. Typical structures
  7. Some hypergraph results
  8. Some (further) important open problems.

  My lectures are planned to be interactive in the sense that the audience can ask questions and I can change the original plans by shifting the originally planned proportions.

  Lecture 4: Estimating the volume and diameter, by Monte Carlo Markov Chain Method

  The results, algorithms of Khachian and later of Karmarkar influenced very much some parts of Theoretical Computer Science. Grötschel, Lovász, and Schrijver wrote an important book [5] on Geometric Algorithms. Under the influence of this book, CO. Elekes [3] proved that the volume of a high dimensional convex body cannot be approximated properly with deterministic algorithms. His results were improved by Bárány and Füredi [1].

  It was an important problem in Theoretical Computer Science to find fast algorithms for some counting problems. L. Valiant showed that counting the 1-factors, i.e. estimating the permanent of a matrix is a “difficult problem”. It was a breakthrough when Jerrum and Sinclair [6] showed that randomized algorithms can approximate the number of 1-factors of a graph. This led to a next breakthrough by Dyer, Frieze and Kannan [2] showing that using randomization the volume can be well approximated. Lovász and Simonovits [9], [10], [11] improved this result, and after a series of papers Lovász and Vempala [12] proved that basically n⁴ oracle questions are enough to approximate the volume. We have to remark that the volume can be approximated, however, the the diameter cannot be approximated.

  In some sense it was natural to approach the area by randomized algorithms, e.g. Monte Carlo methods [14] were fairly popular. The methods mentioned above can also be used in several further areas, see e.g. [4].

  My lecture will give an introduction into this area. Some recommended surveys are the papers of Lovász, and Winkler, e.g. [13], or the homepage of Santosh Vempala [15]

  ---

• Renato Neves (assistant professor at University of Minho, Braga, Portugal)
  Reasoning precisely about imprecisions (metric program equivalence)   toggle abstract   (course details)

  Modern programs increasingly interact with continuous data, randomness, quantum, and physical processes. In this setting, the typical binary notions of program equivalence become too coarse, and more flexible, quantitative ones are called for. In autonomous driving, for example, one does not expect two programs to match the same speeds exactly; instead the idea is to determine how close their behaviours are, typically measured in terms of a suitable metric.

  The course highlights how metric reasoning can be used in practice, with illustrations from probabilistic programming and other computational models. Our focus will be rather on fundamental principles – involving a core programming language (lambda-calculus), corresponding laws of metric equational systems, and their categorical semantics.

  The emphasis throughout is on intuition, the essence of things, and main ideas, rather than full technical detail. Thus no prior knowledge of lambda-calculus or category theory is assumed.

  ---

PROGRAMME

EWSCS is a series of regional-scope international winter schools held annually in Estonia.

The main objective of EWSCS is to expose Estonian, Baltic, and Nordic graduate students in computer science (but also interested students from elsewhere) to frontline research topics usually not covered within the regular curricula. The working language of the schools is English.

EWSCS'26 is the twenty eight event of the series.

To apply for a place, fill in the online participation application form before February 15, 2026.

Admitted participants will be expected to attend all of the school's program. They will be provided access to materials of the courses. Participation includes 3 nights accommodation in twin rooms with full board.

Participation fee: 530 EUR; we may be able to reduce the fee for some BSc and MSc students.

The school will take place at Viinistu Art Hotel. Viinistu is a fishing village in Kuusalu Parish, Harju County in northern Estonia within the territory of Lahemaa National Park. It is located on the coast of the Gulf of Finland on the Pärispea Peninsula by the Eru Bay, about 7 km north of the town of Loksa. From Tallinn the distance is 77 kms.

The school will organize a bus from Tallinn to Viinistu and back.

• Peeter Laud (Cybernetica AS)
• Monika Perkmann (Tallinn University of Technology)
• Pille Pullonen-Raudvere (Cybernetica AS)
• Ago-Erik Riet (University of Tartu)
• Tarmo Uustalu (Reykjavik University / Tallinn University of Technology)

The school is organised by Cybernetica AS. The school is sponsored by the Department of Software Science of Tallinn University of Technology.