diff --git a/README.md b/README.md
index 5e33e3c..ae1b49c 100644
--- a/README.md
+++ b/README.md
@@ -20,6 +20,15 @@ This implementation refers to the project structure of [mulrel-nel](https://gith
 ## Data
 Download [data](https://drive.google.com/file/d/1xW-t80cKDMx3ZL-hrRUxlm6QMZIRvUyU/view) from here and unzip to the main folder (i.e. your-path/DCA). 
 
+This data archive contains the following resource files:
+
+- **Dataset**: One in-domain dataset (AIDA-CoNLL) and Five cross-domain datasets (MSNBC/AQUAINT/ACE2004/CWEB/WIKI). These datasets share the same data format.
+
+- **Type Embedding**: Used to compute type similarity between mention-entity pairs. These type embedding are trained by a typing system called [NFETC](https://arxiv.org/abs/1803.03378) model.
+
+- **Wikipedia inLinks and outLinks**: Surface names of inlinks and outlinks for a Wikipedia page (entity) are used to construct **dynamic context**.
+
+
 ## Installation
 
 Requirements: Python 3.5 or 3.6, Pytorch 0.3, CUDA 7.5 or 8