So I sat down with @Tyr and @geertien to discuss what is AI when it comes to our bots. The initial goal was to further understand the question of “Hard-Coded” vs Machine Learning and how to better classify bots based on their level of complexity and we end up discussing various methodologies in AI, focusing on decision trees, deep learning, and the practical applications in bots.
Highlights
AI Methodologies: The conversation explores different AI methodologies, including deep learning and decision trees. They emphasize that deep learning is just one of many tools in AI development.
Complexity vs. Practicality: Tyr highlights the balance between using advanced AI techniques and simpler hard-coded solutions, noting that while complex AI can be powerful, it’s often more practical to implement straightforward methods.
Search Algorithms for Build Orders: The discussion covers the use of search algorithms to optimize build orders in AI bots, comparing the efficiency of AI-generated solutions to human-devised strategies.
Tech Tree Navigation: Ketroc and Tyr talk about the importance of understanding and implementing tech trees in bot development, ensuring bots can rebuild structures efficiently when necessary.
Game Theory Applications: The episode touches on game theory and its relevance to bot design, providing insights into how AI can be used to predict and counter opponent strategies effectively.