SIGNATURE WORK
CONFERENCE & EXHIBITION 2022

Machine learning has proven itself to be an extremely useful tool in a wide arrange of disciplines. When it’s applied in language processing, it increases the performance of recognizing speeches, when it’s applied in semantic segmentation, now we have cars that can drive on their own. This of course also includes the digital entertainment industry as well. For long, the element of a non-human player (NPC) has always picked the interest of developers and researchers because they serve as a crucial role in the deliverance of a unique experience to the player. But as it has become an integral part of our culture, the expectation from these AIs have increased. The improvements that traditional technics can bring are starting to saturate. Machine learning comes with a potential solution that can fix this up. Through the application this new technology in NPCs, researchers can achieve promising results that show us new potentials with NPC designs with Deep Reinforcement learning, deep neural networks, and many other methods. As such, this paper will aim to provide its readers the general oversight of how machine learning is applied in this area and illustrate current research directions of the field, as well as its limitations.