2026-04-05
00:50 |
AISEE - AI Softwarový expertní vyhledávač pro videa a fotografie
Zitová, Barbara ; Chum, O. ; Smrž, P. ; Šroubek, Filip ; Juránková, M. ; Novozámský, Adam ; Hanousek, Vít ; Bartoš, Michal ; Kamenický, Jan ; Mahdian, Babak ; Průšek, Michal ; Neoral, M.
Souhrnná výzkumná zpráva shrnuje a evaluuje vyvinuté softwarové řešení a jeho chování v okrajových a limitních podmínkách a robustnost při různé kvalitě vstupních videozáznamů. V navazující části obsahuje analýzu použitelnosti a efektivity vybraných AI přístupů využitých v platformě, a to s ohledem na reálné scénáře forenzní praxe. Druhá část se věnuje etickým, právním a lidsko-právním aspektům využití AI při vytěžování dat ve forenzní oblasti a doplňuje doporučení pro zodpovědné používání systému v kontextu aktuální regulace.
Úplný záznam
|
2026-03-01
00:49 |
Statistical Analysis of the 2024 U.S. Presidential Election: Demographics and Swing States
Kalina, Jan
This paper provides an analysis of the 2024 U.S. presidential election using advanced statistical techniques. The study models the popular vote as a response to eight demographic predictors at the state-wide level, incorporating results from the 2020 election to enhance the analysis. A particular focus is given to the application of two recently developed tools inspired by the least weighted squares estimator (LWS): LWS-lasso estimator and LWSquantiles, which are robust methods designed to handle datasets under multicollinearity, heteroscedasticity, and the presence of outliers. The findings emphasize the critical influence of demographic factors in shaping electoral outcomes, illustrating how demographic shifts impact the dynamics of the 2024 election. Special attention is given to the results in seven key swing states, offering precise insights into their pivotal roles in the electoral landscape. Based on the analysis, we propose a novel classification of the swing states into three distinct clusters, taking into account both their demographic outlyingness and their role in the linear model, offering new insights into their strategic importance in the electoral process.
Úplný záznam
|
2026-02-22
00:02 |
Quantum-like Model of a Rat in a Maze
Gaj, Aleksej ; Kárný, Miroslav
Quantum mechanics (QM) provides a formal framework for modelling uncertainty and dynamic evolution in physical systems. While its mathematical structure is well established, its application beyond microscopic phenomena remains an area of active discussion. This work illustrates the axioms of QM in an intuitive, accessible way through a textbook-style example designed to parallel decision-making tasks. The seemingly trivial setup raises questions about the interpretation and under- standing of the underlying model. While it does not aim to introduce new results in quantum theory, it serves as a conceptual bridge, demonstrating fundamental princi- ples in a context involving a living organism. The approach may prove useful as a foundation for quantum-inspired models of decision-making and living systems. Links to contemporary interpretations of QM are also discussed.
Úplný záznam
|
2025-12-15
16:33 |
Structural Learning of BN2A models
Pérez Cabrera, Iván ; Vomlel, Jiří
Bayesian networks are a popular framework for modeling probabilistic relationships between random variables and have been used successfully in educational tests. There is interest in a particular type of Bayesian networks we have called BN2A, which are characterized as bipartite networks, where the first layer consists of hidden variables (which commonly represent skills) and the second layer consists of observed variables (which represent questions in a test). In BN2A models all variables are assumed to be binary. The variables in the second layer depend on the variables in the first layer and this dependence is characterized by conditional probability tables representing Noisy-AND models. In this work, we propose an Expectation-Maximization (EM) algorithm for learning the structure of BN2A models, that is, for learning the relationship between hidden variables and observed variables. To test the structural learning algorithm, we conducted two experiments. For the first experiment, we used synthetic data generated from a BN2A model that we previously defined, while for the second experiment we used a well-known real-world dataset in the field of Cognitive Diagnostic Models, the Fraction Subtraction dataset. Our proposed algorithm has interesting potential use cases. One key application is to generate a reasonably accurate BN2A structure model for educational diagnosis, particularly in scenarios where no prior model exists. Depending on the required level of accuracy, the estimated model can be used directly to analyze skill profiles or serve as an initial framework for test designers, who can further refine it before implementation.
Úplný záznam
|
2025-12-15
16:33 |
Fuzzy Bayesian Networks with Likert Scales
Mrógala, J. ; Perfilieva, I. ; Vomlel, Jiří
Our work is motivated by the applications of probabilistic models in the social sciences, in which surveys and questionnaires are typically used to collect respondents' opinions via a Likert scale. The dividing lines between the states on the Likert scale are vague, so it is natural to interpret them using fuzzy numbers instead of integers. We treat the true model variables as hidden continuous variables, the values of which are observed only through their fuzzified counterparts. This approach seems more conceptually appropriate in the context of surveys and questionnaires, since the modeled variables are continuous by nature but are only observed on a fuzzy, discrete scale. Probabilistic inference with continuous variables is challenging when the assumption of normality of the variables' distribution is violated, which is particularly true for variables modeling polarizing issues. We approximate continuous, multidimensional probability distributions using an F-transform composed of basic functions with central points, called nodes, at a multidimensional grid. We illustrate the suggested approach using a small Bayesian network model of data from the survey ``Dividing Lines in Czech Society.''
Úplný záznam
|
2025-12-15
16:33 |
MDP-Based Analysis of Agent Interactions: From Collaborative to Adversarial Dynamics
Ružejnikov, Jurij ; Guy, Tatiana Valentine
In multiagent systems (MAS), agents often share policy information to influence one another’s decisions. Agent interactions can be categorized as either adversarial or cooperative, and these behaviors can be intentional or unintentional. In the intentional case, agents may share misleading policies to either hinder or support other agents’ decision-making, whereas in the unintentional case, the interaction is merely incidental. From a single-agent perspective, the agent must be able to adapt to various interaction types. This work models MAS using the Multiagent Markov Decision Process (MMDP) and introduces a necessary condition for both intentional cooperative and adversarial interactions. We classify possible policy communications by their truthfulness and intent, and we lay the groundwork for a dynamic, trust-based framework that allows an agent to evaluate and incorporate shared policy information. The proposed approach enables robust and adaptive behaviour in both cooperative and adversarial environments.
Úplný záznam
|
2025-12-15
16:33 |
Výzkumný úkol ČVUT: Tools for Adaptive Portfolio Optimization
Procházka, Tomáš
This research project presents a combination of techniques to develop a sound mathematical approach to the portfolio optimization problem. The problem is formulated as a Linear Quadratic Regulator and solved using Dynamic Programming. The key contributions include integrating multivariate regression modeling of returns with structure estimation for the regressor subset and employing exponential forgetting with an algorithm for varying forgetting factor. The optimal allocation is obtained by solving a constrained quadratic programming problem featuring a custom reward function. We highlight the importance of structure estimation and\nthe sequential approach, while also exploring the potential of modeling optimal allocation using the same regression framework as for returns.
Úplný záznam
|
2025-09-07
02:13 |
2025 DYNALIFE Conference on QUANTUM INFORMATION AND DECISION MAKING IN LIFE SCIENCES PROGRAMME and ABSTRACTS
Guy, Tatiana Valentine ; Kárný, Miroslav ; Gaj, Aleksej ; Ruman, Marko ; Ružejnikov, Jurij ; Pelikán, M.
The DYNALIFE interdisciplinary meeting will explore the role of quantum phenomena in the information processing, transfer of complex living systems, decision making and cognition as well as on methods and techniques for simulating biological informational phenomena.\nThe conference will feature a broad spectrum of topics at the intersection of quantum physics, biology, and information science. We invite submissions of new and original research on a variety of subjects, including but not limited to the following:\n• quantum transport and sensing\n• quantum effects in the brain, perception and cognition\n• role of quantum effects in the origins of life and complexity\n• quantum-to-classical transition in living organisms\n• quantum computing for molecular biology\n• quantum-like modelling and quantum bio inspired technologies\n• quantum decision making and cognition.\nThe conference aims to foster interdisciplinary collaboration among working groups and stimulate cross-field interactions to deepen our fundamental understanding of biological information and its role in the origin of life. Additionally, it will leverage cutting-edge biological insights to describe phenomena in cognition and decision-making and help to explore potential applications (e.g. medicine, technology).
Úplný záznam
|
2025-06-24
13:59 |
How Sir Harold Jeffreys would create a belief function based on data
Daniel, Milan ; Jiroušek, Radim ; Kratochvíl, Václav
Not all normalized nonnegative monotone set functions are belief functions. This paper investigates ways to modify them to obtain a belief function that preserves some of their properties. The problem is motivated by an approach to data-based learning of belief function models. The approach is based on the idea that classical methods of mathematical statistics can provide estimates of lower bounds for unknown probabilities. Thus, methods of mathematical statistics can be used to obtain a reasonable rough estimate, which is further elaborated to obtain a desired belief function model.
Úplný záznam
|
2025-06-24
13:59 |
Úplný záznam
|
|
|
|
host ::
