T2-RAGBench: Text-and-Table Benchmark for Evaluating Retrieval-Augmented Generation
by Kutay Isgörür, Jan Strich, Martin Semmann, Maximilian Trescher, Chris Biemann
Year:
2026
Publication:
arXiv
Abstract:
Since many real-world documents combine textual and tabular data, robust Retrieval Augmented Generation (RAG) systems are essential for effectively accessing and analyzing such content to support complex reasoning tasks. Therefore, this paper introduces T2-RAGBench, a benchmark comprising 23,088 question-context-answer triples, designed to evaluate RAG methods on real-world text-and-table data.
Link:
Read the paperAdditional Information
Brief introduction of the dida co-author(s) and relevance for dida's ML developments.
Kutay Isgörür
While studying Physics at Bogazici University, Kutay focused on theoretical physics and entered the field of machine learning through his involvement in a university research group. He worked on a range of machine learning projects, enhancing his skills in data handling, model training, and analysis. Currently, as a master’s student at Freie University in Berlin, Kutay continues to explore the intersection of computational physics and machine learning. Alongside his studies, he supports the machine learning team at dida.