What is an intelligent fight game coach?
How could I describe it in a few words? A fascinating and engaging stepping stone into artificial intelligence development from my perspective. Here is a summary description of how things are going to work:
- Provide your fighting style preference, for example:
- Zoning: Keep away characters at bay by controlling the distance of your surrounding environment with projectiles and weapons,
- Rushdown: Keep the pressure on the opponent,
- Grappler: Get close and “grapple” with high damage,
- Fight against the artificially-controlled adaptive agent,
- When the fight ends, you get pointers on how to improve by the AI:
- Get highlights of poor and excellent plays made by the human-player during the game,
- Get analyzed weaknesses displayed during the fight,
- Get an update on the summary profile,
- Get recommendations on exercises to go through to improve as a player.
Why would there be a need for such an application?
I am a fighting game enthusiast. I find they provide a nice way to let my competitive nature roam free. My goal is to create an environment in which players can learn and improve. With an AI-coach, players would be able to know how they can become more efficient. This applies to both enthusiasts and professionals. You can find below a list of games you could improve on with such an application:
- The Street Fighter series,
- The Super Smash Bros. series,
- The Tekken series.
With some efforts, it is straightforward to evolve from a rookie to a somehow competent player. From there, it becomes harder to break into the competitive scene and become a recognized authority in the game of your choosing. Without proper guidance, it can quickly become frustrating when a player wants to improve. The goal of the application is to help players get better, by removing the frustration factor.
Career growth opportunities for me
Areas of learning and improvement:
- System design and software architecture
- Cloud computing and AWS
- PyTorch, reinforcement learning, Open AI gym and Unity ML-agents
- Front-end development and React
- Containerization and orchestration
- SQL & NoSQL Databases + Redis
These are areas I’m most interested in and that could have the most impact on my career. This past year, I’ve been committed to challenging myself and growing as an individual contributor. One thing that does make me feel uncomfortable is how this project is getting more tortuous as it goes. I’ve spent some time discussing AWS topics with members of the technical community. I’ve quickly observed that it’s way trickier than I would have expected to put together a design.
What is the project timeline?
Well, I have put together a curriculum of resources to visit and consume. Many topics covered by the application are foreign to me. The project is multi-facet and requires special attention on my end. For now, there is no project timeline. As you can see, the project is indeed intriguing.
There are many areas in which I must throw myself into and ramp up so that I can start building a viable version of this project. A quick estimate shows me I could get something working around mid-year in 2021.
I will be sharing tips on getting started and build a project of that endeavor. It will serve passionate software developers out there that are following this blog.
See you next time,