Qs for Multimodal Dialogue State Tracking by SSNLP 2022 https://padlet.com/ssnlp22/3a_2 SSNLP '22 Session 3A-2 en-us 2022-05-12 14:51:40 UTC 2022-05-12 16:53:52 UTC hello@padlet.com Ask via https://padlet.com/ssnlp22/3a_2 ssnlp22 https://padlet.com/ssnlp22/3a_2/wish/2181408212 2022-05-12 14:52:56 UTC https://padlet.com/ssnlp22/3a_2/wish/2181408212 Is the DVD dataset done on synthetic object videos? ssnlp22 https://padlet.com/ssnlp22/3a_2/wish/2181537964 How do you expect that the model will fare against naturalistic video (e.g., from camera phones)?]]> 2022-05-12 16:09:07 UTC https://padlet.com/ssnlp22/3a_2/wish/2181537964 Would the self-supervised setting also work with naturalistic videos? ssnlp22 https://padlet.com/ssnlp22/3a_2/wish/2181541883 Do the simple types of objects in the dataset lend themselves to higher-fidelity video modeling with VDTN?]]> 2022-05-12 16:11:42 UTC https://padlet.com/ssnlp22/3a_2/wish/2181541883