Seminar by Jan Frąckowiak, PhD Student at the Faculty of Economic Sciences

On 13 April at 17:00, a seminar organized by the QFRG and DSLab research centres will take place. During the event, Jan Frąckowiak, a PhD student affiliated with the Department of Data Science at the Faculty of Economic Sciences, will present the results of his study titled “Ontology as a hyperparameter: Task-aware knowledge graph construction using LLMs”. The presentation will focus on the use of knowledge graphs built from text to create better predictive models. The speaker will also discuss how, by using LLM-based agents, it is possible to automatically evolve the graph structure in response to prediction quality. [The full abstract of the presentation is provided below.]

We cordially invite you to attend the seminar in room B002 or to join online. To receive the link for the online meeting, please contact us at: 2^3SFz5W6yoi|0LM~xeNdp]#[}Q}A2^z5Qq^O#$DVd][WU].

The presentation will last approximately 45 minutes and will be followed by a joint discussion. We kindly ask participants to arrive or log in by 16:50.

---

The project investigates the construction of task-aware knowledge graphs (KGs) from text, treating graph structure as an optimization target rather than a fixed representation. Existing frameworks such as Neo4j-based pipelines and LLM-driven tools (e.g., LlamaIndex) focus on constructing large, general-purpose graphs without incorporating feedback from downstream tasks, despite evidence from evaluation frameworks (e.g., KGrEaT, GEval, KGBench) that KG utility is task-dependent. The research addresses this gap by proposing a feedback-driven approach in which ontology is treated as a tunable component influencing predictive performance. The study uses document-based datasets linking text to measurable outcomes, including financial news (FNSPID) paired with stock price movements. LLM agents are employed for ontology evolution and triple extraction, enabling scalable generation of multiple KG variants. Graph-derived features are then computed over temporal windows and used in predictive models, whose performance is fed back to the agents to refine ontology design, closing the loop between graph construction and downstream task. Preliminary results show that graph structure significantly impacts predictive performance, with adaptive ontology refinement leading to measurable improvements.