@inproceedings{sanabria-etal_VIGIL:2018,
 author = {Sanabria, Ramon and Caglayan, Ozan and Palaskar, Shruti and Elliott, Desmond and Barrault, Loic and Specia, Lucia and Metze, Florian},
 booktitle = {Proceedings of the NeurIPS Workshop on Visually Grounded Interaction and Language},
 title = {How2: A Large-scale Dataset for Multimodal Language Understanding},
 url = {https://arxiv.org/pdf/1811.00347},
 year = {2018}
}