Differences

This shows you the differences between two versions of the page.

--- aira:start [2025/10/28 22:08] – [Schedule Autumn 2025] mzk
+++ aira:start [2025/12/04 16:30] (current) – [2025-12-04] mzk
@@ Line 21: / Line 21: @@
 ===== Schedule Autumn 2025 =====
+  * **[RPHD TRACK] 2025.12.11**: Anna Sofia Lippolis, PhD Candidate @ University of Bologna, [[#section20251211|Enhancing Knowledge Engineering with LLMs.]]
+    * Meeting link:[[https://teams.microsoft.com/meet/35848296892369?p=0t3yTogmgGOsSwYeDO|MS Teams]]
+    * Recording:  TDA
+    * Presentation slides:  TDA
+  * **[PHD TRACK] 2025.12.04**: Bartłomiej Małkus, PhD Candidate @ Jagiellonian University, [[#section20251204|Towards Explainable Meta-Models for Ensembles of Financial Alphas.]]
+    * Meeting link: [[https://teams.microsoft.com/meet/33172349128262?p=yBHWXyoke6oH4OVDl4|MS Teams]]
+    * Recording:  TDA
+    * Presentation slides:  TDA
+  * **[RESEARCH TRACK] 2025.11.27**: Aleksander Mendyk, Professor PhD, DSc @  Jagiellonian University, [[#section20251120|AI/ML for pharmaceutical sciences – an industrial perspective.]]
+    * Meeting link: [[https://teams.microsoft.com/meet/39537797907603?p=jjjJD0gjg5LcGl6eL9|MS Teams]]
+    * Recording:  [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/IQDqGs_UjUSkRIm82hi1-KgBAc4MLc-EjcVPaewZ6sYskI4?e=qTmdDd|View]]
+    * Presentation slides:  {{:aira:slides-aleksander-mendyk-2025-11-27.pdf|Download}}
+  * **[RESEARCH TRACK] 2025.11.13**: Tomáš Kliegr with the research team @ Prague University of Economics and Business, [[#section20251113|RAG research, LLMs as digital twins, Rule Learning in relational data - perspectives in AI Research .]]
+    * Meeting link:[[https://teams.microsoft.com/meet/35466374465161?p=E80GGRrCPxyDKDYomv|MS Teams]]
+    * Recording - Barbara Moreová:  [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/IQBAZ5wBtjiaTYjBlQDQACWHAQFaeBCECRYAmhZiYB1PI6c?e=G4VFdq|View]]
+    * Recording - Mateusz Ploskonka:  [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/IQAC5l947emMQoLcwSSWuqhOATaOnpK3jqxWwR-U-EJ0ODs?e=p2R8tL|View]]
+    * Recording - Kateřina Hrudková:  [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/IQANTKVda1QPS45yKsybdKAYAZ1oSTeQo4b34J7sfF5pbWw?e=DVxlO5|View]]
+    * Presentation slides - Barbara Moreová:  {{:aira:slides-barbara-moreova-2025-11-13.pdf|Download}}
+    * Presentation slides - Mateusz Ploskonka:  {{:aira:slides-mateusz-ploskonka-2025-11-13.pdf|Download}}
+    * Presentation slides - Kateřina Hrudková:  {{:aira:slides-katerina-hrudkova-2025-11-13.pdf|Download}}
+  * **[RESEARCH TRACK] 2025.11.06**: Tomáš Kliegr and Lukáš Sýkora @ Prague University of Economics and Business, [[#section20251106|LLM-based feature generation from text for interpretable machine learning.]]
+    * Meeting link:[[https://teams.microsoft.com/meet/3127857308764?p=1vVlUBWtvXh3pJVm3z|MS Teams]]
+    * Recording - Tomáš Kliegr:  [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/IQA6xvsR6r__Sb0uBw018DxdARL_jxuITdNrnlTomJNFcCI?e=SjdjzJ|View]]
+    * Recording - Lukáš Sýkora:  [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/IQAD1MW0CrT7TbXn7305dHJUAWT51QewwEh9LPyrbfehOj8?e=KfQCx8|View]]
+    * Presentation slides - Tomáš Kliegr:  {{:aira:slides-tomas-kliegr-2025-11-06.pdf|Download}}
+    * Presentation slides - Lukáš Sýkora:  {{:aira:slides-lukas-sykora-2025-11-06.pdf|Download}}
   * **[RESEARCH TRACK] 2025.10.30**: Peter van Dam,  Associate Professor @ Jagiellonian University, [[#section20251030|Inverse problem in electrocardiography: modeling the ECG.]]
     * Meeting link:[[https://teams.microsoft.com/meet/38693113399487?p=QVsurHuK9sLM4PH7ek|MS Teams]]
-    * Recording:  TDA
     * Presentation slides:  TDA
   * **[RESEARCH TRACK] 2025.10.23**: Marek Pędziwiatr,  PhD @ Jagiellonian University, [[#section20251023|Eye tracking as a bridge between psychology and computer science.]]
     * Meeting link:[[https://teams.microsoft.com/meet/3686533128348?p=2ZQTEpXJiiJDSLC3jJ|MS Teams]]
-    * Recording:  TDA
+    * Recording:  [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/IQDdy0LegIdrQZkdzpl-cbfmAcPDZLzhRLLqWoJ4FcsjQp0?e=kGMbFU|View]]
-    * Presentation slides:  TDA
+    * Presentation slides:  {{:aira:slides-marek-pedziwiatr-2025-10-23.pdf|Download}}
   * **[PHD TRACK] 2025.10.16**: Karol Dobiczek,  PhD Candidate @ Jagiellonian University, [[#section20251016|Applying Counterfactual Explanations in Evolving Scenarios and Expert Domains.]]
     * Meeting link:[[https://teams.microsoft.com/meet/3634985739900?p=D3AqLZM4yWIU1nXOWh|MS Teams]]
-    * Recording:  TDA
+    * Recording:  [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/IQD8B4CDbZyuSqIs-5TSzXKSATAQkbiX71yEsz3ogYFMafg?e=hVk2FP|View]]
-    * Presentation slides:  TDA
+    * Presentation slides:  {{:aira:slides-karol-dobiczek-2025-10-16.pdf|Download}}
   * **2025.10.09**: Grzegorz J. Nalepa, AIRAmaster @ Jagiellonian University, //Introduction to AIRA to the new PhD students//
@@ Line 165: / Line 195: @@
       * Meeting link: [[https://teams.microsoft.com/l/meetup-join/19%3aJF1L0935A7eV6s8R9brG5MMYONrqy4XmxPSYLeRMCGM1%40thread.tacv2/1715342946914?context=%7b%22Tid%22%3a%22eb0e26eb-bfbe-47d2-9e90-ebd2426dbceb%22%2c%22Oid%22%3a%22c96e4eee-96f3-4b0f-88ef-fe65310f5f55%22%7d|MS Teams]]
       * Recording: [[https://ujchmura.sharepoint.com/:v:/t/Section_495645_1/EZe0bw0DZY5LtU3CCvns-6UBp432OE_5pD3_vs7zBOPhiA?nav=eyJyZWZlcnJhbEluZm8iOnsicmVmZXJyYWxBcHAiOiJTdHJlYW1XZWJBcHAiLCJyZWZlcnJhbFZpZXciOiJTaGFyZURpYWxvZy1MaW5rIiwicmVmZXJyYWxBcHBQbGF0Zm9ybSI6IldlYiIsInJlZmVycmFsTW9kZSI6InZpZXcifX0%3D&e=eU3MPT|View]] (if you are not UJ employee, ask Szymon Bobek for access)
-      * Presentation slides: {{ :aira:slides-mateusz-bulat-20240515.pdf |Download}}
+      * Presentation slides: {{ :aira:slides-mateusz-bulat-20240515x.pdf |Download}}
   * **[RESEARCH TRACK] 2024.05.09**: Jason J. Jung [[#20240509| Deep Learning for Anomaly Detection in Multivariate Time Series: Approaches, Applications, and Challenges ]]
     * Meeting link: [[https://teams.microsoft.com/l/meetup-join/19%3aJF1L0935A7eV6s8R9brG5MMYONrqy4XmxPSYLeRMCGM1%40thread.tacv2/1715082246704?context=%7b%22Tid%22%3a%22eb0e26eb-bfbe-47d2-9e90-ebd2426dbceb%22%2c%22Oid%22%3a%22c96e4eee-96f3-4b0f-88ef-fe65310f5f55%22%7d|MS Teams]]
@@ Line 466: / Line 496: @@
+==== 2025-12-11 ====
+<WRAP column 15%>
+{{ :aira:anna-lippolis-foto.png?200| }}
+</WRAP>
+<WRAP column 75%>
+**Speaker**: Anna Sofia Lippolis, PhD Candidate @ University of Bologna
+**Title**: Enhancing Knowledge Engineering with LLMs.
+**Abstract**:
+The development and spread of Large Language Models (LLMs) are having a growing impact on the world of the Semantic Web, profoundly transforming the field of Knowledge Engineering. This field, traditionally characterized by a high degree of manual work and collaboration between technical professionals and domain experts, faces various challenges related to scalability and the continuous evolution of knowledge. In this context, LLMs are emerging in several areas, from law to medicine, as tools that support researchers: from the automatic generation of ontologies to the assessment of the quality and semantic coverage of conceptual models, and even the exploration of analogical reasoning, through which it is possible to identify structural correspondences between different domains. This talk will present current research directions on collaboration between LLMs and researchers for knowledge modeling, within a critical overview of the opportunities offered by these tools for the future of the Semantic Web.
+**Biogram**:
+Anna Sofia Lippolis (she/her) is a PhD student at the University of Bologna, Italy, affiliated with the National Research Council’s Institute for Cognitive Sciences and Technologies (Rome, Italy). Her work investigates how semantic technologies intersect with Digital Humanities research and how AI can automate knowledge-engineering practices.
+</WRAP>
+<WRAP clear></WRAP>
+==== 2025-12-04 ====
+<WRAP column 15%>
+{{ :aira:bartlmiej-malkus-foto.jpeg?200| }}
+</WRAP>
+<WRAP column 75%>
+**Speaker**: Bartłomiej Małkus,  PhD Candidate @ Jagiellonian University
+**Title**: Towards Explainable Meta-Models for Ensembles of Financial Alphas.
+**Abstract**:
+One branch of systematic trading research studies large libraries of formulaic alphas: small predictive models built from price and volume data. In practice, these alphas are combined into an ensemble whose composition changes with market conditions. From an ML perspective, this can be viewed as a meta-model that selects and weights weak experts based on their characteristics and the current environment.
+In this talk I will introduce this setting with minimal financial background (cross-sectional returns, information coefficient, long–short factor portfolios), and then reframe it in familiar ML terms. I will show how individual alphas can be treated as models with their own structural and behavioural features, and how this enables clustering them into "families" and reasoning about dynamic ensemble construction. Finally, I will sketch the idea of an explainable meta-model that maps alpha features and market descriptors to ensemble decisions, and highlight open methodological questions and possible research directions.
+**Biogram**:
+Bartłomiej Małkus is a PhD candidate at the Jagiellonian University in Technical Computer Science since 2021. He completed BSc and MSc studies in Computer Science at the AGH University of Science and Technology, and MSc studies in Financial Markets at the Cracow University of Economics. His field of interest are neurosymbolic approaches, prototypical networks, and financial applications of explainable ML. Commercially, he works in IBM on database and master data management solutions.
+</WRAP>
+<WRAP clear></WRAP>
+==== 2025-11-27 ====
+<WRAP column 15%>
+{{ :aira:aleksander-mendyk-foto.png?200| }}
+</WRAP>
+<WRAP column 75%>
+**Speaker**: Aleksander Mendyk, Professor @ Jagiellonian University
+**Title**: AI/ML for pharmaceutical sciences – an industrial perspective.
+**Abstract**:
+Although reluctantly, pharmaceutical industry follows the footsteps of the other industries into the digital age. Among the digital innovations, data-driven approaches are more and more exploited in the various stages of drug discovery and development, including manufacturing as well. AI/ML is perceived as a disruptive technology capable of bringing safe innovation strategies and in the same time extending well known statistical process control and good practices for the benefit of development of safe and efficacious drugs. This movement towards AI/ML applications is stimulated both by industry itself and Regulatory Agencies like EMA and FDA. This talk will outline current areas and trends in the AI/ML applications for the pharmaceutical sciences, which are the backbone of the marketed medicinal products.
+**Biogram**:
+Prof. Aleksander Mendyk, PhD, DSc. is an expert in application of artificial&computational intelligence methods in pharmaceutical technology&biopharmacy, in vitro in vivo correlation (IVIVC) and bioequivalence, Author of over 100 publications. A pharmacist and programmer both in Open Source and commercial applications (R, Python, Java). Currently Head of the Department of Pharmaceutical Technology and Biopharmaceutics, Jagiellonian University-Medical College (JUMC), Kraków, Poland and Vice Dean for Science and Development at the Faculty of Pharmacy JUMC.
+</WRAP>
+<WRAP clear></WRAP>
+==== 2025-11-13 ====
+<WRAP column 15%>
+{{ :aira:Tomáš_Kliegr_research_team-foto.jpg?200| }}
+</WRAP>
+<WRAP column 75%>
+**Speaker**: Tomáš Kliegr with the research team @ Prague University of Economics and Business
+**Title**: RAG research, LLMs as digital twins, Rule Learning in relational data - perspectives in AI Research .
+**Abstract**:
+- Mateusz Ploskonka: RAG Research Presentation - Mateusz will present a comprehensive empirical evaluation of 32 LLMs (from 0.6B to 1T parameters) across various Retrieval-Augmented Generation (RAG) configurations. His research challenges the ""bigger is better"" assumption by identifying a critical performance plateau at 30B parameters, offering evidence-based guidance for practitioners to balance answer quality, cost, and sustainability in production RAG deployments.
+- Barbara Moreová: LLMs as digital twins for simulating effect of cognitive biases - Barbara will present her research on using Large Language Models to create digital twins for the simulation of human cognitive biases.
+- Kateřina Hrudková: Rule Learning in Relational Data: Methods, Interoperability, and Bioscience Applications -  Kateřina’s research bridges the gap between expressive Inductive Logic Programming (ILP) and scalable RDF rule learners. She will present tools (popper2rdf, rdf2popper) that enable this interoperability and demonstrate a practical application using the RDFRules system to discover interpretable ""nuggets"" from large-scale biomedical knowledge graphs, aiding in tasks like drug repurposing for COVID-19 and classifying microbial media.
+**Biogram**:
+Tomáš Kliegr is a Professor at the Faculty of Informatics and Statistics at the Prague University of Economics and Business (VSE Praha), where he is part of the Data Science & Explainable AI (DSXAI) research team. His research interests include Explainable AI (XAI), Interpretable Machine Learning, and neurosymbolic methods. He has published on topics such as the effect of cognitive biases on model interpretation in journals including Artificial Intelligence and Machine Learning. He is active in the rule-based systems community.
+</WRAP>
+<WRAP clear></WRAP>
+==== 2025-11-06 ====
+<WRAP column 15%>
+{{ :aira:kliegr-sykora-foto.jpg?200| }}
+</WRAP>
+<WRAP column 75%>
+**Speaker**: Tomáš Kliegr and Lukas Sykora @ Prague University of Economics and Business
+**Title**: LLM-based feature generation from text for interpretable machine learning.
+**Abstract**:
+Traditional text representations like embeddings and bag-of-words hinder rule learning and other interpretable machine learning methods due to high dimensionality and poor comprehensibility. This article investigates using Large Language Models (LLMs) to extract a small number of interpretable text features. We propose two workflows: one fully automated by the LLM (feature proposal and value calculation), and another where users define features and the LLM calculates values. This LLM-based feature extraction enables interpretable rule learning, overcoming issues like spurious interpretability seen with bag-of-words. We evaluated the proposed methods on five diverse datasets (including scientometrics, banking, hate speech, and food hazard). LLM-generated features yielded predictive performance similar to the SciBERT embedding model but used far fewer, interpretable features. Most generated features were considered relevant for the corresponding prediction tasks by human users. We illustrate practical utility on a case study focused on mining recommendation action rules for the improvement of research article quality and citation impact.
+**Biogram**:
+Tomáš Kliegr is a Professor at the Faculty of Informatics and Statistics at the Prague University of Economics and Business (VSE Praha), where he is part of the Data Science & Explainable AI (DSXAI) research team. His research interests include Explainable AI (XAI), Interpretable Machine Learning, and neurosymbolic methods. He has published on topics such as the effect of cognitive biases on model interpretation in journals including Artificial Intelligence and Machine Learning. He is active in the rule-based systems community.
+Dr. Lukas Sykora is a Research Assistant at the Department of Information and Knowledge Engineering and a Lecturer at the Prague University of Economics and Business.  He holds a PhD in Applied Informatics (2025), where his doctoral thesis focused on action rule mining. He has authored several publications on this topic, including "Apriori Modified for Action Rules Mining." He also brings industry experience as a Solution Architect Team Lead at Ogilvy.
+</WRAP>
+<WRAP clear></WRAP>
+==== 2025-10-30 ====
+<WRAP column 15%>
+{{ :aira:peter-vanDam-foto.jpg?200| }}
+</WRAP>
+<WRAP column 75%>
+**Speaker**: Peter van Dam,  Associate Professor @ Jagiellonian University
+**Title**: Inverse problem in electrocardiography: modeling the ECG.
+**Abstract**:
+The electrocardiogram (ECG) measured on the body surface shows beat-by-beat the electrical functioning of the heart. This ECG consists of a number of electrical time signals that are caused by the currents produced by the myocardial cells. These ECG signals require clinical interpretation, which in the In current clinical practice is close to an art. Modeling the ECG, i.e. have a model of the heart that simulates the electrical processes of the myocardium and calculates the ECG signals on the body surface, can help to understand these waveforms in a cause -effect relationship. This is the major subject of my current and past research, which I will introduce during this talk.
+**Biogram**:
+Dr hab. Peter van Dam is a scientist and lecturer specializing in cardiac modeling, electrocardiographic diagnostics, and the development of innovative educational and clinical tools based on heart signal analysis and 3D imaging. In 2025, he obtained his postdoctoral degree (habilitation) in medical sciences at the Jagiellonian University Medical College, presenting a scientific achievement entitled “Localization of PVC (Premature Ventricular Contraction): Evaluation and Validation of Inverse ECG Modeling.” He is also a laureate of the prestigious NCN OPUS grant, focused on inverse modeling of repolarization heterogeneity to assess arrhythmogenic risk.
+His academic path began with studies in electronic engineering at MBO-College Gouda (1981–1985) and Hogeschool Utrecht (1985–1990). He also studied philosophy (Rijksuniversiteit Utrecht, 1991–1992) and physics (Universität Oldenburg, 1992–1993). Between 1997–2004, he earned a master’s degree in biology and physics at the Open University in Heerlen, where his thesis involved developing a simulation of the P-wave in ECG and IEGM (2004). In 2010, he defended his PhD at Radboud University Nijmegen under the supervision of Prof. Adriaan van Oosterom, presenting the dissertation “The shortest path to cardiac activation.”
+Peter van Dam’s professional experience spans both academia and the medical industry. He began his career in 1997 as a senior scientist at Vitatron/Medtronic. Since 2009, he has been affiliated with Peacs BV (Chief Scientific Officer). In 2020, he became CEO of ECG-Excellence, focusing on developing innovative diagnostic and educational tools based on heart signal analysis and 3D modeling.
+In academia, he has worked at Radboud University (2011–2017, validating ECGI methods in an animal model), at UCLA (2013–2016, developing arrhythmia mapping methods), and currently conducts research and teaching at UMC Utrecht, UT Twente, and Hochschule Luzern Technik & Architektur. Since 2023, he has also collaborated with the Jagiellonian University Medical College, Center for Digital Medicine and Robotics, where he serves as Visiting Professor at the Department of Radiology. Since 2024, he has also been a specialist at the Laboratory of Functional and Virtual Medical 3D Imaging (3D-FM) at the University Hospital in Kraków.
+His research focuses on: inverse ECG modeling and arrhythmia source localization; the use of 3D imaging and simulation in cardiology; developing modern educational tools for students and physicians; integrating engineering and clinical expertise in the diagnosis of cardiac rhythm disorders.
+Dr hab. Peter van Dam sees great potential in interdisciplinary collaboration between universities and university hospitals worldwide. His goal is to create and implement groundbreaking 3D ECG technologies that not only enable more precise diagnostics but also open new possibilities for more effective treatment of cardiac arrhythmias. By combining clinical and engineering perspectives, he strives to set new standards in modern cardiology.
+</WRAP>
+<WRAP clear></WRAP>
 ==== 2025-10-23 ====