A Dataset To Evaluate The Representations Learned By Video Prediction Models
Published in
International Conference on Learning Representa...(2018)External Links:
TL
TL;DR
This work presents a parameterized synthetic dataset called Moving Symbols to support the objective study of video prediction networks and proposes the use of a performance metric with greater semantic meaning to improve experimental interpretability.
Abstract
We present a parameterized synthetic dataset called Moving Symbols to support the objective study of video prediction networks. Using several instantiations of the dataset in which variation is explicitly controlled, we highlight issues in an existing state-of-the-art approach and propose the use of a performance metric with greater semantic meaning to improve experimental interpretability. Our dataset provides canonical test cases that will help the community better understand, and eventually improve, the representations learned by such networks in the future. Code is available at this https URL .
References12 items
1
Decomposing Motion and Content for Natural Video Sequence Prediction
2
Unsupervised Learning of Disentangled Representations from Video
3
Video Pixel Networks
4
Learning a Driving Simulator
5