# Generative Temporal Models with Spatial Memory for Partially Observed Environments

arXiv: https://arxiv.org/abs/1804.09401 · HTML: https://mlsanity.com/paper/1804.09401/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 Apr 2018.

## Built on by (most cited first, 10 of 10)

- [Unconstrained Scene Generation with Locally Conditioned Radiance Fields](https://mlsanity.com/paper/2104.00670/) (2021)
- [On the link between conscious function and general intelligence in humans and machines](https://mlsanity.com/paper/2204.05133/) (2022)
- [Spatially-Aware Transformer for Embodied Agents](https://mlsanity.com/paper/2402.15160/) (2024)
- [Kanerva++: extending The Kanerva Machine with differentiable, locally block allocated latent memory](https://mlsanity.com/paper/2103.03905/) (2021)
- [Tracking and Planning with Spatial World Models](https://mlsanity.com/paper/2201.10335/) (2022)
- [Continual Learning of Multi-modal Dynamics with External Memory](https://mlsanity.com/paper/2203.00936/) (2022)
- [Building spatial world models from sparse transitional episodic memories](https://mlsanity.com/paper/2505.13696/) (2025)
- [A Biologically-Inspired Dual Stream World Model](https://mlsanity.com/paper/2209.08035/) (2022)
- [PRISM: Probabilistic Real-Time Inference in Spatial World Models](https://mlsanity.com/paper/2212.02988/) (2022)
- [ContraNeRF: 3D-Aware Generative Model via Contrastive Learning with Unsupervised Implicit Pose Embedding](https://mlsanity.com/paper/2304.14005/) (2023)

## Abstract (arXiv)

In model-based reinforcement learning, generative and temporal models of environments can be leveraged to boost agent performance, either by tuning the agent's representations during training or via use as part of an explicit planning mechanism. However, their application in practice has been limited to simplistic environments, due to the difficulty of training such models in larger, potentially partially-observed and 3D environments. In this work we introduce a novel action-conditioned generative model of such challenging environments. The model features a non-parametric spatial memory system in which we store learned, disentangled representations of the environment. Low-dimensional spatial updates are computed using a state-space model that makes use of knowledge on the prior dynamics of the moving agent, and high-dimensional visual observations are modelled with a Variational Auto-Encoder. The result is a scalable architecture capable of performing coherent predictions over hundreds of time steps across a range of partially observed 2D and 3D environments.