Releasing illustrative colab for ensemble_loss_landscape.

clarahu · derpson · commit 7dc61f1deb3a · 2020-12-16T12:52:37.000Z
PiperOrigin-RevId: 346832894
diff --git a/README.md b/README.md
@@ -59,6 +59,8 @@ https://deepmind.com/research/publications/
 *   [Graph Matching Networks for Learning the Similarity of Graph Structured
     Objects](graph_matching_networks), ICML 2019
 *   [REGAL: Transfer Learning for Fast Optimization of Computation Graphs](regal)
+*   [Deep Ensembles: A Loss Landscape Perspective](ensemble_loss_landscape)
+
 
 
 
diff --git a/ensemble_loss_landscape/README.md b/ensemble_loss_landscape/README.md
@@ -0,0 +1,12 @@
+# Accompanying code for Deep Ensemble: A Loss Landscape Perspective
+
+[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/deepmind/deepmind_research/blob/master/ensemble_loss_landscape/cifar10_medium_cnn_experiments.ipynb)
+
+The Colab notebook `cifar10_medium_cnn_experiments.ipynb` illustrates the CIFAR-10
+experiments in the paper:
+
+[Deep Ensembles: A Loss Landscape Perspective](https://arxiv.org/abs/1912.02757)
+by Stanislav Fort, Huiyi Hu and Balaji Lakshminarayanan
+
+These experiments investigate the effects of ensembling and variational Bayesian
+methods, please see the paper for more details.
diff --git a/ensemble_loss_landscape/cifar10_medium_cnn_experiments.ipynb b/ensemble_loss_landscape/cifar10_medium_cnn_experiments.ipynb
