As a starter of the information he will share from the digital stage of SIGGRAPH 2020, the programmer and researcher Daniel Holden of Ubisoft Montreal shows the ambitious Deep Learning technology that it is developing to make animations in nextgen games even more realistic and natural.
The technology in question, explains Ubisoft Montreal expert, will be based on Machine Learning and the latest advances made in the Neural AI to reduce the amount of memory used by traditional Motion Matching techniques.
In addition to the documentation that summarizes the benefits and features of this new development system of Deep Learning animations, we also find short explanatory videos that testify to the enormous potential of this technology.
Using this system, Holden reports that he can automatically and effortlessly create a huge database of animations, all using the computational power and artificial intelligence of a Neural network. In this way, reports the Ubisoft author, it will be possible to shape nextgen video games with characters, creatures, animals and abstract beings with extremely realistic animations, above all capable of reacting and adapting to the surrounding environment.
Before leaving to Holden's tweet with the demonstration movie on Motion Matching in Deep Learning of the French videogame giant, we remind you that Ubisoft Montreal collaborated with 14 studios to create Assassin's Creed Valhalla and, at the end of 2019, a mysterious game was canceled after three years of development.