News of the Institute of Computer Science Polish Academy of Sciences

On January 31, an article by Prof. Wojciech Penczek, Łukasz Maśko and Teofil Sidoruk, written in collaboration with Prof. Laure Petrucci, Carlos Olarte, and Jaime Arias from Université Sorbonne Paris Nord, was published in IEEE Transactions on Reliability. The paper "Optimal Scheduling of Agents in ADTrees: Specialized Algorithm and Declarative Models" follows earlier work [1] that proposed representing attack-defence trees (ADTrees) as multi-agent systems. ADTrees are a popular formalism for studying security scenarios where two opposing parties attempt to achieve sub-goals (tree nodes) comprising the overall objective (tree root), or try to prevent the other group from doing so. With the translation to an agent-based formalism, one can additionally consider the attackers and defenders in these scenarios as agent coalitions. As such, not only do they have a particular number of agents, but also specific assignment to sub-goals. This can determine not only the feasibility of an attack or defence, but also its time, cost, or other quantitative metrics of interest.

A Polish large language model PLLuM, trained on mainly Polish-language content, and an intelligent assistant based on it will be developed by six Polish research units. 'We cannot afford to be left behind,’ say the project’s representatives.

The PLLUM (Polish Large Language Universal Model) consortium was established in November (on the eve of chatGPT's first birthday). Its members are six of Poland's leading research units in the field of artificial intelligence and linguistics: the Wrocław University of Science and Technology (consortium leader), NASK National Research Institute, National Information Processing Institute - National Research Institute, the Institute of Computer Science of the Polish Academy of Sciences, the University of Łódź and the Institute of Slavic Studies of the Polish Academy of Sciences. Representatives of NASK announced the establishment of the consortium in a press release sent to PAP.

The Statistical Analysis and Modeling Group of IPI PAN achieved significant results regarding two new machine learning methods for positive-unlabeld data in an important case when their availability depends on the characteristics of the examined entities. The first proposed method examined a new parametric model for generating this type of data and an iterative procedure for estimating its parameters. The second method is based on the concept of variational autoencoders combined with outlier detection techniques. In addition to the theoretical justification of the methods, a significantly better quality of the resulting classifiers for tabular and image data was experimentally demonstrated.

Both methods were presented at the European Conference on Artificial Intelligence, ECAI 2023, which took place in Krakow on September 30-October 4, 2023.

In an article accepted for this year's Annual Meeting of the Association for Computational Linguistics – the most important and largest conference on computational linguistics – a professor of linguistics from the Institute of Computer Science of the Polish Academy of Sciences and the University of Warsaw, together with a student of cognitive science at the University of Warsaw – presented a new argument for symmetric theories of coordinate structures.

This argument is based on automatic processing of large collections of texts, in this case English texts, and on advanced statistical analysis of extracted linguistic data.