Meta's Motivo AI model

Meta has only released Meta Motivo, a new artificial intelligence model built to make avatars move and look more realistic in the Metaverse. The social media giant has not always received applause for avatars, and according to the company, this discovery can enable its avatar agents to be realistic in their movements and could be the first step towards ‘creating full immersive agents within the Metaverse.’

Meta’s Motivo AI Model

Meta demonstrated a brief 50-second video on X of how this new model would work with the use of virtual avatars. Meta claims it is using ‘unsupervised reinforcement learning’ to cause these ‘human-like behaviors’ in animated characters to make them seem as real as possible. The clip demonstrated a humanoid avatar executing dance moves and kicks by whole-body tasks that emerged from a mathematical model.

According to the company, this research breakthrough allows its virtual agents to have more natural and realistic movements and could pave the way for creative agents within the Metaverse. Despite small talk and limited generalization, the proposed model can output human-like behaviors and perform comparably to task-specific methods and is superior to the state-of-the-art unsupervised RL and model-based baselines.

The company insists it has what it calls Motivo as the first-of-its-kind behavioral foundation model. The AI model can also solve many challenging whole-body tasks for the Virtual Human Avatar and make virtual physics in the Metaverse more realistic.

Banking on investments in AI-augmented reality and other Metaverse resources, it has been pumping tens of billions of dollars into its investments and has thus increased capex projections for 2024 to as much as between $37-$40 billion, far from the previous 2019 level of $26 billion. Meta Motivo responds to such body control issues affecting digital avatars, hence improving how these avatars march in specific movements in a natural, human-like way, according to the company.

Meta’s Video Seal

In addition to Motivo, Meta also released Meta Video Seal, an open-source tool for adding AI-driven watermarks to videos. It offers the possibility to embed a completely invisible watermark within the video, which could be used to prove the origin and thus, the authenticity of the video, significantly helping to deal with fake news or manipulated content.

To More Connect with us Follow us on Facebook, Twitter, Pinterest, and Reddit. Stay updated!

By Goldy Choudhary

Goldy Choudhary serves as the Manager of Clinic Beauty Store in Raleigh, North Carolina, USA where she leverages AI tools such as Lumen5, ChatGPT, and Gemini to drive innovation and enhance operational efficiency. With a deep-seated passion for the AI revolution, Goldy contributes to AyuTechno as a part-time author, where she plays a crucial role in content creation.Her commitment to the field of artificial intelligence is evident through her daily experiences and research, which she translates into valuable content for AyuTechno. Goldy’s role is instrumental in providing readers with comprehensive insights into AI and guiding them on secure and effective usage of these technologies. Her mission is to empower individuals with knowledge and ensure they are well-informed about the latest advancements in AI, reflecting her dedication to making AI accessible and safe for all.

Leave a Reply

Your email address will not be published. Required fields are marked *