Visual Attention in Imaginative Agents
arXiv:2104.00177
Abstract
We present a recurrent agent who perceives surroundings through a series of discrete fixations. At each timestep, the agent imagines a variety of plausible scenes consistent with the fixation history. The next fixation is planned using uncertainty in the content of the imagined scenes. As time progresses, the agent becomes more certain about the content of the surrounding, and the variety in the imagined scenes reduces. The agent is built using a variational autoencoder and normalizing flows, and trained in an unsupervised manner on a proxy task of scene-reconstruction. The latent representations of the imagined scenes are found to be useful for performing pixel-level and scene-level tasks by higher-order modules. The agent is tested on various 2D and 3D datasets.
References in corpus (9)
- NICE: Non-linear Independent Components Estimation
- Recurrent Models of Visual Attention
- DRAW: A Recurrent Neural Network For Image Generation
- Multiple Object Recognition with Visual Attention
- Nonparametric variational inference
- Emergence of Exploratory Look-Around Behaviors through Active Observation Completion
- On Learning Where To Look
- Block Neural Autoregressive Flow
- Anticipating many futures: Online human motion prediction and synthesis for human-robot collaboration