Search in AI 3D Environments

This is my second post on <a href="https://k3no.medium.com/3d-virtual-environments-for-ais-e02de5c4ef1d" rel="noopener">3D Environments for AIs</a>, which aim to provide a safe and convenient place to test and generate AIs. The first post covered much of the background and setup, but we barely scratched the surface. This post will delve deeper into actually giving a task for our AIs to do. — — — — — — — — — — —  <code><a href="https://medium.com/membership/@k3no" rel="noopener">SUBSCRIBE TO MEDIUM !</a></code>  — — — — — — — — — — <h2>Why search ?</h2> Search is considered a now classic and mostly solved (though not perfected) problem or task for AIs. “AI: A Modern Approach” (AIMA), often regarded as the AI Bible, devotes most of its early chapters to search, for instance. <pre> Why 3D ? At first glance there's not much to be gained from implementing search in 3D as it adds an extra step for visualization purposes. However, I chose to take this route to provide both contrast and a foundation for setting up more complex AI tasks. As mentioned earlier, there is a strong connection to video games where search (or pathfinding) is actually relevant. Additionally, I am also learning as I write this.</pre> For us humans and other animals search and Intelligence go hand in hand as we are always in search of something, food, shelter, money, love, a missing sock, you name it and most of what we call intelligence deals with finding that right something. While you might associate search with data search engines like Google, LLMs like ChatGPT can also be considered to be doing search ( searching the next relevant tokens to continue a conversation ). <a href="https://k3no.medium.com/search-in-ai-3d-environments-ad2b0b7c4de1">Visit Now</a>