Initial release of himo.

PiperOrigin-RevId: 337916498

README.md

@@ -24,6 +24,7 @@ https://deepmind.com/research/publications/
 
 ## Projects
 
+*   [Value-driven Hindsight Modelling](himo), NeurIPS 2020
 *   [Targeted free energy estimation via learned mappings](learned_free_energy_estimation), Journal of Chemical Physics 2020
 *   [Learning to Simulate Complex Physics with Graph Networks](learning_to_simulate), ICML 2020
 *   [Physically Embedded Planning Problems](physics_planning_games)

himo/README.md

@@ -0,0 +1,18 @@
+# Accompanying code for Value-driven Hindsight Modelling
+
+`himo_example.ipynb` contains code for the illustrative task of Section 3.2 in the [paper](https://arxiv.org/abs/2009.05524).
+
+## Requirements and Installation
+
+This is a Colaboratory notebook: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/deepmind/deepmind_research/blob/master/himo/himo_example.ipynb)
+
+If your runtime doesn't include [Haiku](https://github.com/deepmind/dm-haiku) and [Tree](https://github.com/deepmind/tree) please run the cell labeled "Install extra dependencies" which contains the following pip installation commands:
+
+```!pip install dm-haiku```
+
+```!pip install dm-tree```
+
+
+## Reference
+**Value-driven Hindsight Modelling** Arthur Guez, Fabio Viola, Théophane Weber, Lars Buesing, Steven Kapturowski, Doina Precup, David Silver, and Nicolas Heess
+Neural Information Processing Systems 2020, URL: https://arxiv.org/abs/2009.05524