diff --git a/README.md b/README.md
index 4a445af..c42f03a 100644
--- a/README.md
+++ b/README.md
@@ -7,35 +7,34 @@ Implemented algorithms:
 * Deep Q-Learning (DQN)
 * Double DQN
 * Dueling DQN
+* Async Advantage Actor Critic (A3C)
 * Async One-Step Q-Learning
 * Async One-Step Sarsa 
 * Async N-Step Q-Learning
-* Async Advantage Actor Critic (A3C)
 
 # Curves
 > Curves for CartPole is trivial so I didn't place it here.
-## Deep Q-Learning (DQN)
-![alt text](https://raw.githubusercontent.com/ShangtongZhang/DeepRL/master/images/DQN-BreakoutNoFrameskip-v3-Train.png)
-![alt_text](https://raw.githubusercontent.com/ShangtongZhang/DeepRL/master/images/DQN-BreakoutNoFrameskip-v3-Test.png)
+## DQN, Double DQN, Dueling DQN 
+![Loading...](https://raw.githubusercontent.com/ShangtongZhang/DeepRL/master/images/DQN-breakout.png)
+![Loading...](https://raw.githubusercontent.com/ShangtongZhang/DeepRL/master/images/DQN-Pong.png)
 
-The network and parameters here are exactly same as the DeepMind Nature paper. 
-Training curve is smoothed by window of size 100. Test is triggered every 1000 episodes.
-In total it took about 16M frames. Training time is 4 days and 10 hours in a server with
-Xeon E5-2620 v3 and Titan X.
+The network and parameters here are exactly same as the [DeepMind Nature paper](https://www.nature.com/nature/journal/v518/n7540/full/nature14236.html). 
+Training curve is smoothed by a window of size 100. All the models are trained in a server with
+Xeon E5-2620 v3 and Titan X. For Breakout, test is triggered every 1000 episodes with 50 repetitions.
+In total, 16M frames cost about 4 days and 10 hours. For Pong, test is triggered 
+every 10 episodes with no repetition. In total, 4M frames cost about 12 hours.
 
-## Asynchronous Advantage Actor Critic (A3C)
+## A3C
 
-![alt text](https://raw.githubusercontent.com/ShangtongZhang/DeepRL/master/images/A3C-PongNoFrameskip-v3.png)
+![Loading...](https://raw.githubusercontent.com/ShangtongZhang/DeepRL/master/images/A3C-PongNoFrameskip-v3.png)
 
 The network I used here is same as the network in DQN except the activation function 
 is **Elu** rather than Relu. The optimizer is **Adam** with non-shared parameters.
 To my best knowledge, this network architecture is not the most suitable for A3C. 
-If you use a 42 * 42 input, add a LSTM layer, you will get much much much better training speed 
+If you use a 42 * 42 input, add a LSTM layer at last, you will get **much much much** better training speed 
 than this. [GAE](http://www.breloff.com/DeepRL-OnlineGAE/) can also improve performance.
-Another important thing is I didn't use lock for syncing up networks. Although I think there
-should be a lock, locking can hurt the performance heavily (about 50%). 
 
-The first 15 million frames took about 5 hours (16 processes) in a server with two Xeon E5-2620 v3.
+The first 15M frames took about 5 hours (16 processes) in a server with two Xeon E5-2620 v3.
 This is the test curve. Test is triggered in a separate deterministic test process every 50K frames.
 
 # Dependency
@@ -52,5 +51,7 @@ Detailed usage and all training details can be found in ```main.py```
 * [Asynchronous Methods for Deep Reinforcement Learning](https://arxiv.org/abs/1602.01783)
 * [Deep Reinforcement Learning with Double Q-learning](https://arxiv.org/abs/1509.06461)
 * [Dueling Network Architectures for Deep Reinforcement Learning](https://arxiv.org/abs/1511.06581)
+* [Playing Atari with Deep Reinforcement Learning](https://arxiv.org/abs/1312.5602)
+* [HOGWILD!: A Lock-Free Approach to Parallelizing Stochastic Gradient Descent](https://arxiv.org/abs/1106.5730)
 * [transedward/pytorch-dqn](https://github.com/transedward/pytorch-dqn)
 * [ikostrikov/pytorch-a3c](https://github.com/ikostrikov/pytorch-a3c)