Functionality can be defined as “the particular use for which an object is designed”. For example, the functionality of a cart is to allow humans to store and move products in a supermarket. Moreover, computers have been commonly used for designing 3D objects in a semi-automatic manner, where the human user provides high-level constraints for a design, while the computer defines the fine geometric details of the object based on the constraints. Given that many manufactured objects have a specific functionality, it is a natural goal to also seek the development of computer tools that can design objects with a required functionality. The technical challenges for accomplishing this goal are the development of a method that can discover and evaluate the functionality of an object in an automatic manner, and a method to involve the user in the design process in an interactive manner.
In this talk, I will discuss some of our work towards the goal of designing functional objects in a semi-automatic manner. First, I will discuss a line of work where we automatically evaluate aspects of the functionality of objects based on an analysis of the objects and use of machine learning methods. Specifically, we learn to predict the functionality of objects by training a model of functionality from data containing examples of object functionality. Then, I will show an example of a system that uses the functionality model to semi-automatically design objects that possess a certain functionality. Finally, I will discuss challenges and future directions in the design of objects based on functionality, especially involving the use of deep learning methods for shape modeling and how a user can be involved in the design process.
AI systems in digital games have been developed in these 30 years, and it converged to MCS-AI dynamic cooperative model (Meta-Character-Spatial AI dynamic cooperative model). The model consists of three AIs such as Meta AI, Character AI and Spatial AI. Meta-AI is to control a whole game situation by giving orders to NPCs (non-player characters) and objects in a game. A character AI is a brain of an NPC to make a decision by itself. Spatial AI is to analyze a terrain in a game and inform them to the other AIs. The model is also effective for a smart city. Meta-AI controls a whole city situation, A character AI is a brain of each drone or robot, and spatial AI is to analyze a situation of the city in real-time. The model produces services to people and protects a city against disasters and accidents. In the session, the design of MCS-AI dynamic cooperative model and some cases in a game are explained.