paint-brush
How AI Helps to Make the SIDUS Metaverse Betterby@dankhomenko
106 reads

How AI Helps to Make the SIDUS Metaverse Better

by Dan KhomenkoMarch 20th, 2023
Read on Terminal Reader
Read this story w/o Javascript
tldt arrow

Too Long; Didn't Read

SIDUS has been involved in AI-tech game development for over 15 months. The team actively uses the toolset by Topaz Labs for image and video quality enhancement. AI-assisted tools help us to make key poses very fast, see physical results instantly, and adjust secondary motion that allows us to triple speed of animation creation.

People Mentioned

Mention Thumbnail
featured image - How AI Helps to Make the SIDUS Metaverse Better
Dan Khomenko HackerNoon profile picture

For over 15 months, SIDUS has been involved in AI-tech game development. During 2022 only, Artificial Intelligence (AI) and Machine Learning (ML) helped us optimize our game development, and in the future, we are going to push for further optimization.

Along with the decrease in expenses, AI and ML help us to boost the quality of every component of the SIDUS Metaverse. 

AI and ML Solutions Used by the Sidus Heroes Team

We follow the newest and most promising AI and ML developments and apply them to make the SIDUS Metaverse better and to boost the gaming experience of our users. 

The SIDUS HEROES team actively uses the toolset by Topaz Labs for image and video quality enhancement. We believe that this is the best toolset for a concept artist. 

Cascadeur is tool that we love and use. This AI-assisted keyframe animation software helps us to create ultra-realistic animations. 

AI-assisted tools help us to make key poses very fast, see physical results instantly, and adjust secondary motion that allows us to triple the speed of animation creation and boost its quality.

To create expressive facial animations from an audio source, the team uses Nvidia’s Audio2Face. It allows us to create an animated 3D face with just an audio clip! 

Another Nvidia AI solution we love and plan to use in our open world is a new AI for simulations.

Even the simplest computer simulation requires huge volumes of data to be stored. So, for some simulations, hundreds of gigabytes of data will be stored for a simulation that lasts just a couple of minutes. Neural networks have exceptional compression capability that allows to reduce the volumes of data stored without compromising video quality. We plan to use this solution for open world development. 

Currently, we are working on tasks to make real-time animation possible. We are already at the testing stage for our characters, and once this stage is over, this technology will allow us to move our media to a new quality level.

We Keep an Eye on Emerging AI Solutions

Our team is also busy with RnD technologies that are not available as ready-made tools. Take this one, for example. 

We believe that this technology is going to revolutionize the approach to game animation for perfectly seamless character motion. 

A big part of our RnD is developing bots for all genres. Xenna uses generative bots and for our upcoming third-person shooter, we are researching bots that use ML to achieve the most intelligent and lifelike behavior. Our bots get new data input after each battle and become increasingly lifelike. The more battles that are completed, the smarter the bots become. 

Generative Algorithms and ML Algorithms Are Not the Same

Machine Learning algorithms are often mistaken for ordinary generative algorithms. We use and admire the first ones. We love this tool and are planning to use it for creating open world locations. 

This tool allows us to simulate a physics and biology-based ecosystem. With it, vegetation is the function of precipitation and temperature, just like in the real world. If there is no rain and low temperature, we get a tundra-like landscape. With no rain and high temperature, a desert-like landscape is generated

Not All ML and AI Algorithm Are Acceptable

Even though it is undeniable that AI and ML solutions facilitate our work immensely, there are tools that we never use due to ethical reasons. Neural networks that generate images are becoming increasingly popular among designers. But we believe that the application of such neural networks as DALEE or OpenArt contradicts our idea of creativity, and using the images generated by such tools is nothing more than theft. These networks undermine the concept of an artist as somebody who delivers new ideas. 

Why?

Because whatever these types of neural networks do is not unique. It is a sublimation of works created by artists previously. There is nothing really creative or original in such works.

Our works are 100% unique and created by our imagination and our understanding of the Sidus Metaverse lore. 

Final Thoughts

AI and ML algorithms are irreplaceable when it comes to developing a virtual world. We are using the most innovative of them for the development of all aspects of the SIDUS Metaverse, and we are going to use them to provide our users with unique unforgettable user experiences.