linagora/MinecraftStructuredDialogueCorpus

Minecraft Structured Dialogue Corpus

The Minecraft Structured Dialogue Corpus (MSDC) is a discourse-annotated version of the Minecraft Dialogue Corpus (MDC), first created by Julia Hockenmaier's lab in 2019.

The MSDC features complete, situated discourse structures for each dialogue in the style of Situated Discourse Representation Theory (Asher and Lascarides, 2003).

Dataset Description

The MSDC is a collection of natural language dialogues between pairs of human participants who meet in a simulated 3D Minecraft grid and communicate through a chat window. One assumes the role of Architect, and the other the role of Builder. At the beginning of each dialogue, the Architect is shown an image of a structure made from colored blocks, and the goal of the interaction is to help the Builder recreate the structure on the grid via typed Architect instructions.

The chat interaction between the Architect and Builder produces linguistic speaker turns, while the Builder actions (the placement and removal of blocks) are represented textually, e.g. place orange 2 1 -1, and attributed to Builder as speaker.

A small group of discourse experts and students annotated the corpus. First they hand-segmented the linguistic turns into elementary discourse units (EDUs). Each Builder action turn was considered a single elementary event unit (EEU). All but 7 of the dialogues in the MDC were annotated, and the original MDC splits were preserved, except for a small Development set culled from Test for the purpose of developing the Llamipa discourse parser.

	Train	Val	Test	Dev	Total
No. Dialogues	307	100	101	32	540
No. EDUs	9616	3297	3009	1128	17050
No. EEUs	6669	2240	2006	732	11647
No. Relation instances	19598	6681	6018	2232	34529

For more corpus stats and an in-depth description of the corpus creation, please see the paper below.

Citations

Please cite the paper below if using the MSDC in your work.

Kate Thompson, Julie Hunter and Nicholas Asher, 2024. Discourse Structure for the Minecraft Corpus. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 4957–4967, Torino, Italia. ELRA and ICCL.

@inproceedings{thompson-etal-2024-discourse,
    title = "Discourse Structure for the {M}inecraft Corpus",
    author = "Thompson, Kate  and
      Hunter, Julie  and
      Asher, Nicholas",
    editor = "Calzolari, Nicoletta  and
      Kan, Min-Yen  and
      Hoste, Veronique  and
      Lenci, Alessandro  and
      Sakti, Sakriani  and
      Xue, Nianwen",
    booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)",
    month = may,
    year = "2024",
    address = "Torino, Italia",
    publisher = "ELRA and ICCL",
    url = "https://aclanthology.org/2024.lrec-main.444/",
    pages = "4957--4967"
}

Acknowledgements

We acknowledge support from the National Interdisciplinary Artificial Intelligence Institute, ANITI (Artificial and Natural Intelligence Toulouse Institute), funded by the French ‘Investing for the Future–PIA3’ program under the Grant agreement ANR-19-PI3A-000. We also thank the ANR project COCOBOTS (ANR-21-FAI2-0005), the ANR/DGA project DISCUTER (ANR21-ASIA-0005), and the COCOPIL “Graine” project funded by the Région Occitanie of France.