Update README.md

Author: MichaelYxWang

README.md

@@ -12,6 +12,9 @@ Project Manager: [Taylor Zhang](https://www.linkedin.com/in/xingzhi-taylor-zhang
 
 This is our Final Year Design Project in Electrical and Computer Engineering Department at University of Waterloo.
 
-Our team is researching a scenery-music recommender system under the supervision of [Dr. Alexander Wong](https://www.eng.uwaterloo.ca/~a28wong/). Designed for the visually impaired, users can start recording a short video of their ambient surroundings. Next, our application extract frames of the video, and then the image captioning model would generate a description of the surroundings. Finally, we recommend suitable songs based on the similarity between the generated captions and the potential lyrics. Recently, we had a successful [DEMO (Video)](https://u.nu/demo-cnib) (ML demo starts at 10:00 in the video) with the Canadian National Institute for the Blind. Sincerely, we believe that our design philosophy to connect various domains – CV, NLP, and mobile development – with our “AI for Good” initiative blessed us with success.
+Our team is researching a scenery-music recommender system under the supervision of [Dr. Alexander Wong](https://www.eng.uwaterloo.ca/~a28wong/). Designed for the visually impaired, users can start recording a short video of their ambient surroundings. Next, our application extract frames of the video, and then the image captioning model would generate a description of the surroundings. Finally, we recommend suitable songs based on the similarity between the generated captions and the potential lyrics. Recently, we had a successful [Presentation Session (Video)](https://u.nu/demo-cnib) (ML demo starts at 10:00 in the video) with the Canadian National Institute for the Blind. Sincerely, we believe that our design philosophy to connect various domains – CV, NLP, and mobile development – with our “AI for Good” initiative blessed us with success.
 
-For Machine Learning related code, please check "ml-research" and "ml-prod" directories.
+
+We also attached a small demo video in the repo (demo.mp4). In this demo, we first select a picture with snow and upload it to our mobile app. Then the algorithm generates both the captions and recommended music. You can see we get pretty accurate prediction and recommendation!
+
+If you want to walk through Machine Learning related code, please check "ml-research" directory.