About
Jakub holds BSc and MSc in Computer Science and graduated from Wroclaw University of Science and Technology. His interest in machine learning and artificial intelligence starts in the early year of his studies. He has experience in scientific and commercial projects. During his career, he developed and deployed solutions in several ML fields, including natural language processing (automated anonymization, fake news classification) and signal processing (acoustic scene classification). His current research concerns learning representation on graphs and its applications to problems, like embedding neural network learning processes.
Research Interests
Education
MSc in Computer Science
Wroclaw University of Science and Technology · 2021
BSc in Computer Science
Wroclaw University of Science and Technology · 2020
Publications
Dissecting Mahalanobis: How Feature Geometry and Normalization Shape OOD Detection
2025 · arXiv 2510.15202
FactSelfCheck: Fact-Level Black-Box Hallucination Detection for LLMs
2025 · arXiv 2503.17229
Hallucination Detection in LLMs Using Spectral Features of Attention Maps
2025 · EMNLP 2025
The Illusion of Progress: Re-evaluating Hallucination Detection in LLMs
2025 · EMNLP 2025
Empowering Small-Scale Knowledge Graphs: Leveraging General-Purpose KGs for Enriched Embeddings
2024 · LREC-COLING 2024
Enhancing Out-of-Distribution Detection Through Stochastic Embeddings in Self-supervised Learning
2024 · ICCS 2024
A Deeper Look at Graph Embedding RetroFitting
2023 · J. Computational Science 68
RAFEN: Regularized Alignment Framework for Embeddings of Nodes
2023 · ICCS 2023
Graph-Level Representations Using Ensemble-Based Readout Functions
2023 · ICCS 2023
Fact-checking: relevance assessment of references in the Polish political domain
2021 · *Knowledge-Based and Intelligent Information & Engineering Systems: Proceedings of the 25th International Conference KES 2021*
