From 30452cf1b6fa14ecfa39720ac86ff2d9f1b365da Mon Sep 17 00:00:00 2001
From: Tian Lan <31748898+Emerald01@users.noreply.github.com>
Date: Thu, 1 Aug 2024 22:59:41 -0700
Subject: [PATCH] Update README.md

---
 example_envs/rlchemists/README.md | 18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

diff --git a/example_envs/rlchemists/README.md b/example_envs/rlchemists/README.md
index bc22a4c..84e9f57 100644
--- a/example_envs/rlchemists/README.md
+++ b/example_envs/rlchemists/README.md
@@ -1,6 +1,6 @@
 # rlchemists
 
-This is the source code folder for the research project published on Nature Communications ...
+This is the source code folder for the research project published on Nature Communications: https://www.nature.com/articles/s41467-024-50531-6/figures/3
 
 ## structure
 1. `en_array`: energy landscape mesh from DFT
@@ -12,8 +12,24 @@ This is the source code folder for the research project published on Nature Comm
 ## installation
 1. setup GPU environment and install `warpdrive` package as instructed
 2. under the root directory of `rlchemists`, run `bash setenv.sh` to setup the Python path for this project
+
+## CUDA kernel for environment
+Please contact the authors for the kernel functions as this is not OSS.
    
 ## run
 We simply choose the environment and type to run a particular training, the supported ones are all included
 in the `run_configs` folders, for example, `run_configs/single_agent_one_atom_diffusion2d` can be run by 
 `python example_training_script_numba.py --env single_agent_one_atom --type diffusion2d`
+
+## cite
+If you're using this study in your research or applications, please cite using this BibTeX:
+```
+@article{lan2024,
+  title  = {Enabling high throughput deep reinforcement learning with first principles to investigate catalytic reaction mechanisms.},
+  author = {Lan, Tian and Wang, Huan and An, Qi},
+  year   = 2024,
+  journal = {Nature Communications},
+  volume  = {15},
+  number  = {6281},
+}
+```