Disclaimer: This article may contain some outdated information.

OORT (previously Computecoin) hosted an AMA event with the founder of Uverse, John Jiang, yesterday at 9 AM in OORT’s official Telegram chat. Yesterday marks a significant milestone for both OORT and Uverse — Uverse, an AI-powered Avatar platform for the metaverse, is the brand new addition to the OORT Ecosystem.

Note: CCN is the testnet token of OORT's first (Dome-A) and second (Huygens) testnets.

Hi John, thanks for joining us today. Can you tell us a bit about yourself and Uverse?

Hi everyone. I’m Johnson Jiang. You can call me John. I’m very honored and excited to be here and chat with you all. My academic background is in geophysics and linguistics, and my specialties are in machine translation, NLP, and multi-modality AI engine research and development. I’ve built several companies and held multiple patents in these fields.

Uverse is the world’s first multi-modality AI chatbot, audio simulation, and video generation platform. Uverse’s advanced AI technology enables users to generate video content featuring lifelike digital avatars that intelligently emulate the way people speak, including the mechanics of speech, as well as verbal intonation and other factors.

What differentiates Uverse from other digital human/AI avatar solutions on the market right now?

Uverse is founded on several key innovations. First, Uverse’s AI improves its computational precision through multilingual, rather than probabilistic training. Its linear semantic analysis system also enables Uverse to deepen its understanding of natural linguistic patterns.

Uverse also uses GPU for its computation, allowing it to organize every component of a sentence and organize translations through multi-hop attention, glimpses, and gating.

Uverse “comprehends” and translates text through contextualized, multilayered natural language processing. Uverse’s comparative computations are founded on multi-granular generative adversarial networks.

Uverse can produce audio content that features diverse dynamic characteristics by deciphering nuanced emotional content in the text analysis stage and during speech emotion recognition.

Next, Uverse generates 3D digital content in multiple languages that accurately expresses emotion both orally and visually. These innovations make Uverse more intelligent, more efficient, and more accurate than other solutions on the market.

What are some of the critical advantages of Uverse?

Uverse is multilingual, multi-modality, and multigranular. Our text-to-audiovisual solution supports up to 55 languages; replicates various subtle communication cues, including unique vocal timbre, pronunciation and expression, mood, pauses, and many others; and generates highly detailed digital people with distinct face shapes, complexions, facial expressions, and more.

More generally, Uverse will make it possible to generate live chat, audio, and video content through a computer without ever needing actors, cameras, and microphones. Once it’s fully developed, our solution will boast all the advantages of an AI engine, including infinite scalability, virtually zero marginal costs, and wide accessibility; everyone will be able to create high-quality video content with Uverse.

What challenges have you faced in developing Uverse?

Developing our highly sophisticated metaverse AI algorithms involved new territories of AI technologies. Text interaction is based on the individual human soul. We wanted to build a solution to create a pure simulacrum of personalized human life instead of its mere environments, which was a steep challenge.

Designing our product to operate in multiple languages proved difficult because each language has its characteristics and nuances of reflecting human thoughts and even its unique history and cultural heritage.

The supply of computing power is also a challenge because our new AI algorithms often require millions of steps and thousands of epochs to train a model. Thankfully, we found the best computing solution for our dilemma: the innovative OORT infrastructure.

Why OORT? Many cloud platforms are available to choose from, including AWS, Google Cloud, Azure, etc.

Uverse will become a core asset in the metaverse because our architecture effectively brings digital humans to life and hosts them in the virtual realm for real humans to interact with. We chose to partner with OORT because their commitment to laying the cornerstone of the metaverse resonated with our vision of Uverse playing an indispensable role in the future of the internet and humanity’s digital experience.

OORT delivers a massive supply of rich, fast, and diverse computing power. That means OORT can help Uverse offer a high-fidelity solution to users, generally improving the user experience and allowing Uverse to widely deploy our solution without being burdened by excessive lag, fidelity, and scaling problems. OORT made our scalability easy — we need 1000 machines to work simultaneously to train either ten avatars or 1000 avatars without much difference.

Uverse needs extensive and low-cost computing power for AI training. OORT Network integrates tremendous computing power worldwide, making it possible for us to shop around for high-reliable and low-cost computing resources. The computing resource provided by OORT has been very reliable — the cost is at least 30%-50% off compared to that of AWS.

What applications are digital humans uniquely suited for?

Digital humans can step in for real humans to perform time-intensive tasks and engage with people in settings where rich verbal communication is essential.

Education is an area where we believe digital humans will thrive. We offer four functional solutions that meet the needs of students, including textbook recitation, real-time video textbooks, AI teacher Q&A, and expert services. Uverse’s textbook recitation solution will keep students engaged with the material and shorten students’ reading time to minimize eye strain. Our video textbook solution will enable textbook publishers to save on production costs and time.

Virtual teachers will mimic the student’s real teachers at school and provide the student with a personable and engaging studying experience. Uverse’s AI teacher Q&A solution will simulate the student’s real teacher’s facial expressions and unique vocal characteristics, helping to lessen teachers’ burden by providing 24-hour, high-quality education for students. Finally, our expert service will learn answers to questions from our expert team and store the answers in a database, over time accumulating a vast repository of information for students to consult at any time.

Where can we expect to interact with digital humans first in our daily lives?

We will eventually interact with digital humans in various domains, the first of which are likely to include education, retail, sales, entertainment, and healthcare.

What are the factors driving digital human adoption right now?

Digital humans address a salient need: Users benefit from more personalized, more personable service, and demand for this high-quality service far outpaces supply. In recent years, the conventional approach to this challenge has been to implement chatbots. While effective in some ways, chatbots fall short in interpreting the nuances of human verbal and non-verbal communication (causing them to misunderstand users). They lack the familiarity and approachability of a human face, which users find more compelling than a chatbot.

However, the adoption of digital humans also faces the challenge of cost and the limited availability of high-quality, affordable computing power.

Where is Uverse a leader in the race to develop lifelike, intelligent digital humans?

Our solution (once it’s fully developed) will not only support over fifty languages but will also understand, capture, and replicate the fine-grained details of human expression, both verbal and facial, making our communication so rich in information and emotion.