%0 Conference Proceedings %T From Dataset Recycling to Multi-Property Extraction and Beyond %A Dwojak, Tomasz %A Pietruszka, Michał %A Borchmann, Łukasz %A Chłędowski, Jakub %A Graliński, Filip %Y Fernández, Raquel %Y Linzen, Tal %S Proceedings of the 24th Conference on Computational Natural Language Learning %D 2020 %8 November %I Association for Computational Linguistics %C Online %F dwojak-etal-2020-dataset %X This paper investigates various Transformer architectures on the WikiReading Information Extraction and Machine Reading Comprehension dataset. The proposed dual-source model outperforms the current state-of-the-art by a large margin. Next, we introduce WikiReading Recycled - a newly developed public dataset, and the task of multiple-property extraction. It uses the same data as WikiReading but does not inherit its predecessor’s identified disadvantages. In addition, we provide a human-annotated test set with diagnostic subsets for a detailed analysis of model performance. %R 10.18653/v1/2020.conll-1.52 %U https://aclanthology.org/2020.conll-1.52 %U https://doi.org/10.18653/v1/2020.conll-1.52 %P 641-651