Week 1 First Impression

We were so excited to join the CMU-SV campus in the beautiful EA office. Our project is to experiment with reinforcement learning, a relatively new machine learning method, to create believable game characters. Based on our topic, we chose our team name as “Shadow Agent.” There were two meaning behind the name. One is that we are the controller behind the scene, hidden like a shadow. The second meaning is that our game characters would share the same interface and behavior as the real player character, acting similarly like a shadow.

Besides choosing the team name, we met our client, Erin, for the first time. Erin gave us a detailed explanation of the project and talked about her expectation towards the end of the semester. With her information, we began exploring the possible game genres and existing reinforcement learning applications in games.

Team meeting

On Friday, August 30th, we had a wondering BBQ bonding time at Carl’s house, with a beautiful yard and delicious food.

BBQ at Carl’s house

We are now ready to share what we have found to our clients and talk more about the technical limitations of reinforcement learning. We are excited to meet the whole team at Google for our second client meeting.