Datasets:

dasyd
/

quants

@@ -204,37 +204,37 @@ size_categories:
 QuAnTS is a challenging dataset designed to bridge the gap in question-answering research on time series data.
 The dataset features a wide variety of questions and answers concerning human movements, presented as tracked skeleton trajectories.
-To benchmark the practical usability of models trained on this dataset, QuAnTS also includes human performance metrics.
-**For more information, please refer to the paper (Under Review):** *TODO*
-At present, there is no official leaderboard for this dataset.
-## Supported Tasks
-The primary task for the QuAnTS dataset is **Question Answering**. Given a time series of human skeleton trajectories and a question in natural language, the goal is to generate a correct answer.
-## Languages
 The text in the dataset is in English.
 ## Licensing, Citation, and Acknowlegements
 The QuAnTS dataset is licensed under the [Creative Commons Attribution 4.0 International (CC BY 4.0)](ttps://creativecommons.org/licenses/by/4.0/ ) license.
-## Dataset Curators
 The dataset was curated by a team of researchers from various institutions:
-*   Felix Divo, Maurice Kraus, and Kristian Kersting (hessian.AI, DFKI, and the Centre for Cognitive Science) from TU Darmstadt.
 *   Anh Q. Nguyen, Hao Xue, and Flora D. Salim from UNSW Sydney.
 *   Imran Razzak from Mohamed bin Zayed University of Artificial Intelligence.
 *   Devendra Singh Dhami from Eindhoven University of Technology.
-If you use the QuAnTS dataset in your research, please cite:
-*TODO*
-### Acknowlegements
-TODO

 QuAnTS is a challenging dataset designed to bridge the gap in question-answering research on time series data.
 The dataset features a wide variety of questions and answers concerning human movements, presented as tracked skeleton trajectories.
+To benchmark the practical usability of models trained on this dataset, QuAnTS also includes human reference performance.
+![Q: "What is the person doing first?", A: "They are waving.", Q: "How many times are they jumping after that?", A: "..."](doc/intro-chat.png "Example chat motivating time series question answering")
+**For more information, please refer to the paper:** *Under Review*
+At present, there is no official leaderboard for this dataset.
+## Task and Format
+The primary task for the QuAnTS dataset is Time Series Question Answering.
+Given a time series of human skeleton trajectories and a question in natural language, the goal is to generate a correct answer.
+Answers are provided in one of the following formats binary (Yes/No), multiple-choice (A/B/C), or open (free text).
+Additionally, to provide more training data for free text answers, we provide fully textual answers for binary and multiple-choice, too.
+The ground truth action sequence or scene descriptions *may not* be used for answering the dataset — they are provided for debugging purposes.
 The text in the dataset is in English.
+We provide fixed splits into training, validation, and test portions, where only the latter may be used to compare performance across different approaches.
+You are free to mix the training and validation splits as needed.
 ## Licensing, Citation, and Acknowlegements
 The QuAnTS dataset is licensed under the [Creative Commons Attribution 4.0 International (CC BY 4.0)](ttps://creativecommons.org/licenses/by/4.0/ ) license.
+### Dataset Curators
 The dataset was curated by a team of researchers from various institutions:
+*   Felix Divo, Maurice Kraus, and Kristian Kersting (hessian.AI, DFKI, and the Centre for Cognitive Science) from Technische Universität Darmstadt.
 *   Anh Q. Nguyen, Hao Xue, and Flora D. Salim from UNSW Sydney.
 *   Imran Razzak from Mohamed bin Zayed University of Artificial Intelligence.
 *   Devendra Singh Dhami from Eindhoven University of Technology.
+If you use the QuAnTS dataset in your research, please cite: *Under Review, TODO*