Update README.md

README.md

@@ -6,14 +6,20 @@ pipeline_tag: reinforcement-learning
 tags:
 - sac
 - rl-agent
-- tensorflow.js
+- tensorflow-js
 - javascript
-- reinforcement learning
-- computer vision
-- Soft-Actor Critic
+- reinforcement-learning
+- computer-vision
+- soft-actor-critic
 ---
-# ai-creature.github.io
+# The AI Creature
 
 The AI Creature is a Reinforcement Learning agent based on the Soft-Actor Critic algorithm. The agent is capable of navigating and interacting within its environment, utilizing Binocular Vision to distinguish between different stimuli. This system, built with TensorFlow.js and Convolutional Neural Networks, shows remarkable learning efficiency, achieving significant milestones in computer vision tasks with minimal training time.
 
-Demo: [https://www.youtube.com/watch?v=y1GWcVLL-Xw](https://www.youtube.com/watch?v=y1GWcVLL-Xw)
+The training of 5 neural networks (1 actor, 2 critics, 2 target critics) and inference happen on the client side in the browser.
+
+It takes the agent 1-2 hours of trial and error to learn to avoid red spheres and pursue green ones (the browser tab must remain in focus to prevent the browser's optimization algorithm from stopping training when the tab is idle).
+
+**Website:** [https://ai-creature.github.io](https://ai-creature.github.io)
+
+**Demo Video**: [https://www.youtube.com/watch?v=y1GWcVLL-Xw](https://www.youtube.com/watch?v=y1GWcVLL-Xw)