Petra Vidnerová, née Kudová Department of Artificial Intelligence Institute of Computer Science, The Czech Academy of Sciences Pod Vodárenskou věží 2, 182 07 Praha 8, Czech Republic phone: (+420) 266 053 630 petra cs.cas.cz [GitHub] [Twitter] [Research Gate] [Google Scholar]

---

Research interests:

Machine learning, supervised learning, kernel methods, regularization networks. Deep learning.
RBF networks (RBF Layer for Keras).
Hyper-parameter setup, metalearning. Neural architecture search.
Genetic algorithms, evolutionary and hybrid approaches.

Multi-agent network models.

Projects:

TS01020123	Optimalizace energetického portfolia pro zvýšení využitelnosti obnovitelných zdrojů	TA ČR	2024 - 2026 (PI)
25-15490S	LEDNeCo: Low Energy Deep Neurocomputing	Czech Grant Agency	2025 - 2027 (team member)
	TRUST	OP JAK	2025 - 2028 (activity leader)
	DEZINFO	OP JAK	2025 - 2028 (team member)
	AI: Artificial Intelligence for Science and Society	Strategie AV21	2025 - 2029 (package coordinator)

---

News:

• [26. 7. 2025] Paper J. Šíma, P. Vidnerová: Cross-Entropy Loss of Approximated Deep Neural Networks has been accepted for ICONIP 2025


• [27. 5. 2025] Paper J. Šíma, P. Vidnerová: Weight-Rounding Error in Deep Neural Networks has been accepted for ECML PKDD 2025


• [26. 5. 2025] Talk at ICS: Neural Architecture Search: Mapping the field via network text analysis and my journey through the field


• [8. 4. 2025] Přednáška na Kick-off meetingu projektu Epicentra civilizace, Strategie AV21: Modelování rodinných interakcí pomocí generativních agentů


• [27. 2. 2025] Přednáška na Kick-off meetingu projektu TRUST, VA3: Síťová analýza jako nástroj studia racionality textu / vědeckého poznání.


• [6. 11. 2024] Popularizační přednáška na dni otevřených dveří ÚI, P. Vidnerová, G. Kadlecová Co se skrývá za AutoML?.


• [19. 9. 2024] Two working papers presented as posters at SBP-BRiMS 2024.
  G. Kadlecová, P. Vidnerová, R. Neruda, J.Šlerka. Creating Artificial Survey Panels Is Still Difficult ,
  P. Vidnerová, G. Kadlecová, R. Neruda, J. Šlerka. Modelling Information Spread using Generative Agents .


• [17. 9. 2024] A conference paper J. Kalina, P. Vidnerová: On the Bayesian Interpretation of Robust Regression Neural Networks at the ICANN 2024 conference.


• [26. 6. 2024] Talk Information Spread Modelling Using Generative Agents at CoRE conference, Prague.


• [30.5.2024] The paper J. Šíma, J. Cabessa, P. Vidnerová On energy complexity of fully-connected layers was accepted for publication in Neural Networks.


• [2. 5. 2024] Paper G. Kadlecová, J. Lukasik, M. Pilát, P. Vidnerová, M. Safari, R. Neruda, F. Hutter Surprisingly Strong Performance Prediction with Neural Graph Features has been accepted to ICML 2024 .


• [10. 4. 2024] The join talk with G. Kadlecová Modelling Pitfalls and What Can We Get from LLMs at CoRE Seminar at FF UK.


• [19. 3. 2024] The join talk with G. Kadlecová Performance Prediction for NAS at Hora Informaticae Seminar.


• [17. 3. 2024] The paper J. Šíma, P. Vidnerová, V. Mrázek: Energy Complexity of Convolutional Neural Networks was accepted for publication in Neural Computation.


• [12. 1. 2024] Přednáška Multi-agentní epidemiologické modely při příležitosti návštěvy předsedkyně AV ČR na ÚI.


• [9. 1. 2024] The paper J. Kalina, P. Vidnerová, P. Janáček Highly robust training of regularized radial basis function networks was accepted for publication in the journal Kybernetika.


• [24. 11. 2023] The presentation The 2022 Election in the United States: How to Verify Reliability of Linear Regression at RELIK 2023 conference.


• [9. 11. 2023] The presentation Performance prediction for Neural Architecture Search + poster at DaiZ'23 meetup.


• [6. 11. 2023] The paper C. Brom, T. Diviák, J. Drbohlav, V. Korbel, R. Levínský, R. Neruda, G. Suchopárová, J. Šlerka, M. Šmíd, J. Trnka, P. Vidnerová: Rotation-based schedules in elementary schools to prevent COVID-19 spread: A simulation study published in Scientific Reports.


• [26. 9. 2023] Presentation Properties of the weighted and robust implicitly weighted correlation coefficients at ICANN 2023 conference.


• [24. 9. 2023] Two conference papers in ICANN 2023 conference proceedings:
  J. Šíma, P. Vidnerová, V. Mrázek: Energy complexity model for convolutional neural networks.
  J. Kalina, P. Vidnerová: Properties of the Weighted and Robust Implicitly Weighted Correlation Coefficients


• [8. 9. 2023] Presentation Neural Networks - Energy and Robustness at Jizerka Off-site Seminar.


• [7. 9. 2023] The paper Importance of vaccine action and availability and epidemic severity for delaying the second vaccine dose was awarded The Best Result of ICS in 2022 in the category Paper with Application or Social Impact.. Presentation at Jizerka Off-site Seminar.


• [23. 5. 2023] Přednáška o modelu M na křtu monografie Rok s pandemií covid-19.


• [28. 3. 2023] Paper: L. Berec, T. Diviák, A. Kuběna, R. Levínský, R. Neruda, G. Suchopárová, J. Šlerka, M. Šmíd, J. Trnka, V. Tuček, P. Vidnerová, M. Zajíček:
  On the contact tracing for COVID-19: A simulation study .
  Published in Epidemics .


• [13. 2. 2023] Vyšla monografie Rok s pandemií covid-19 (Reflexe v poločase) .


• [6. 12. 2022] Talk at the seminar Hora Informaticae: Model M - an agent-based epidemiological model .


• [22. 7. 2022] Presentation at IJCNN conference (WCCI 2022):
  J. Kalina, P. Vidnerová, P. Janáček: Sparse Versions of Optimized Centroids .


• [24. 6. 2022] Conference paper: G. Suchopárová, P. Vidnerová, R. Neruda, M. Šmíd:
  Using a Deep Neural Network in a Relative Risk Model to Estimate Vaccination Protection for COVID-19
  Published in Conference proceedings of EANN 2022.


• [21. 6. 2022] Paper: L. Berec, J. Smyčka, R. Levínský, E. Hromádková, M. Šoltés, J. Šlerka, V. Tuček, J. Trnka, M. Šmíd, M. Zajíček, T. Diviák, R. Neruda, P. Vidnerová:
  Delays, Masks, the Elderly, and Schools: First Covid-19 Wave in the Czech Republic.
  Published in Bulletin of Mathematical Biology .


• [20. 6. 2022] Presentation at ICAISC 2022 conference:
  P. Vidnerová, J. Kalina Multi-objective Bayesian Optimisation for Neural Architecture Search


• [10. 5. 2022] Paper: L. Berec, R. Levínský, J. Weiner, M. Šmíd, R. Neruda, P. Vidnerová, G. Suchopárová: Importance of vaccine action and availability and epidemic severity for delaying the second vaccine dose.
  Published in Scientific Reports .


• [10. 5. 2022] R. Neruda's talk Tested on agents - how we designed an agent-based epidemiological model


• 2021 - 2019

---

• Curriculum Vitae (pdf) [last update: March, 2025]
• PhD Thesis

• List of Publications
• Slides from presentations

• HOWTOs
• Bookmarks

---

---