Datasets:
license: cc-by-4.0
NLU Evaluation Data - English and German
A labeled English and German language multi-domain dataset (21 domains) with 25K user utterances for human-robot interaction.
This dataset is collected and annotated for evaluating NLU services and platforms. The detailed paper on this dataset can be found at arXiv.org: Benchmarking Natural Language Understanding Services for building Conversational Agents
The dataset builds on the annotated data of the xliuhw/NLU-Evaluation-Data
repository. We have added an additional column (answer_de)
by translating the texts in column answer into German.
The translation was made with DeepL.
Copyright
Copyright (c) the authors of xliuhw/NLU-Evaluation-Data
Copyright (c) 2022 Philip May
All data is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0).