CrossTask is a dataset for studying weakly supervised learning through cross-tasks in instructional videos. Each task provides an ordered list of steps with a manual description. The dataset contains instructional videos collected for 83 different tasks, divided into two parts:

• 18 main tasks: manually collect and provide annotations of temporal step boundaries;
• 65 related tasks: automatically collected, without annotations

Citation

@INPROCEEDINGS{Zhukov2019,

author = {Zhukov, Dimitri and Alayrac, Jean-Baptiste and Cinbis, Ramazan Gokberk and Fouhey, David and Laptev, Ivan and Sivic, Josef}, title = {Cross-task weakly supervised learning from instructional videos}, booktitle = CVPR, year = {2019}, }