1. Predictable artificial intelligenceLexin Zhou, Pablo A. M. Casares, Fernando Martínez-Plumed, John Burden, Ryan Burnell, Lucy Cheke, Cèsar Ferri, Alexandru Marcoci, Behzad Mehrbakhsh, Yael Moros-Daval, Danaja Rutar, 2026, original scientific article Abstract: Many areas of artificial intelligence, and machine learning in particular, aim at being probably correct, i.e., valid on average, rather than pursuing the idealistic goal of being provably valid for all inputs. However, AI systems could still be predictably valid, such as an imperfect robot deliverer for which we can reliably and precisely predict the task instances for which it is correct and safe, its valid operating range. “Predictable AI” is a nascent research area that explores ways of anticipating key validity indicators (e.g., performance, safety) of present and future AI ecosystems. We argue that achieving predictability is crucial for fostering trust, liability, control, alignment and safety of AI, and thus should be prioritised over performance. We formally characterise predictability, explore its most relevant components, illustrate what can be predicted, describe alternative candidates for predictors, as well as the trade-offs between maximising validity and predictability. To illustrate these concepts, we bring an array of illustrative examples covering diverse ecosystem configurations. “Predictable AI” is related to other areas of technical and non-technical AI research, but have distinctive questions, hypotheses, techniques and challenges. This paper aims to elucidate them, calls for identifying paths towards a landscape of predictably valid AI systems and outlines the potential impact of this emergent field.
Keywords: predictable AI, general-purpose AI, AI safety
Published in RUP: 09.02.2026; Views: 255; Downloads: 5
 Full text (4,86 MB)
This document has more files! More... |
2. Generalization of edge general position problemPaul Manuel, R. Prabha, Sandi Klavžar, 2025, original scientific article Abstract: The edge geodesic cover problem of a graph G is to find a smallest number of geodesics that cover the edge set of G. The edge k-general position problem is introduced as the problem to find a largest set S of edges of G such that at most k-1 edges of S lie on a common geodesic. We show that these are dual min-max problems and connect them to an edge geodesic partition problem. Using these connections, exact values of the edge k-general position number is determined for different values of k and for various networks including torus networks, hypercubes, and Benes networks.
Keywords: general position set, edge geodesic cover problem, edge k-general position problem, torus network, hypercube, Benes network
Published in RUP: 03.11.2025; Views: 337; Downloads: 3
Full text (812,56 KB) |
3. On some extremal position problems for graphsJames Tuite, Elias John Thomas, Ullas Chandran S.V., 2025, original scientific article Abstract: The general position number of a graph G is the size of the largest set of vertices S such that no geodesic of G contains more than two elements of S. The monophonic position number of a graph is defined similarly, but with `induced path' in place of `geodesic'. In this paper we investigate some extremal problems for these parameters. Firstly we discuss the problem of the smallest possible order of a graph with given general and monophonic position numbers. We then determine the asymptotic order of the largest size of a graph with given general or monophonic position number, classifying the extremal graphs with monophonic position number two. Finally we establish the possible diameters of graphs with given order and monophonic position number.
Keywords: general position, monophonic position, Turán problems, size, diameter, induced path
Published in RUP: 21.10.2025; Views: 364; Downloads: 2
Full text (396,13 KB) |
4. Coverings of general digraphsAleksander Malnič, Boris Zgrablić, 2025, original scientific article Abstract: A unified theory of covering projections of graphs and digraphs is presented as one theory by considering coverings of general digraphs, where multiple directed and undirected edges together with oriented and unoriented loops and semiedges, are allowed. It transpires that coverings of general digraphs can display certain pathological behaviour since the naturally defined projections of their underlying graphs may not be coverings in the usual topological sense. Consequently, homotopy does not always lift, although the unique walk lifting property still holds. Yet, it is still possible to grasp such coverings algebraically in terms of the action of the fundamental monoid. This action is permutational and has certain nice properties that monoid actions in general do not have. As a consequence, such projections can be studied combinatorially in terms of voltages. The problem of isomorphism and equivalence, and in particular, the problem of lifting automorphisms, is treated in depth. All known results about covering projections of graphs are simple corollaries of just three general theorems.
Keywords: mixed graph, general digraph, dart, covering projection, voltage, homotopy, monoid action, lifting automorphisms
Published in RUP: 10.09.2025; Views: 675; Downloads: 18
Full text (569,42 KB) |
5. |
6. Attitudes towards school, self-perceived school competence and general self-esteem during and after the 1st wave of COVID-19 epidemic in Slovenia : a case studyVesna Posavčević, 2021, original scientific article Keywords: case study, COVID-19 epidemic, distance schooling, elementary school, children, development, attitudes towards school, self-perceived school competence, general self-esteem
Published in RUP: 24.06.2021; Views: 2561; Downloads: 17
 Link to full text |
7. |
8. |
9. |
10. |
