Less than two years later, Facebook’s parent company Meta has announced plans to go “all in” to virtual reality and the Metaverse. The company has recently focused its efforts on Generative AI, a hot topic in the tech world right now, as consumer engagement on these two initiatives has proved disappointing so far.
Generative AI refers to a trending class of machine learning applications that can create new data, such as text, images, video, and sound, based on large-scale trained datasets. Examples of generative AI applications include ChatGPT. fastest growing application The same is true for image creation tools such as Dall-E and Stable Diffusion.
Experts now predict that this technology will disrupt every industry, impacting the products and services we consume and the way we work. So here’s a look at some of the ways Meta has implemented these powerful tools across the platform, as well as how it impacts their ongoing plans to launch us all into the Metaverse. Here are some ideas.
generative advertising
Facebook, the largest platform on Meta and the world’s largest social network, makes money primarily by allowing businesses to advertise on its pages.now Said The company announced that it will provide generative AI tools to these companies as the first commercialization of its proprietary generative AI technology.
Later this year, the company is expected to release a tool that will allow the company to automate the creation of multiple versions of ads featuring different text and images for different audiences. Automate elements such as language used, colors and even celebrities and influencers featured in promotions to appeal to different groups of people depending on their age, interests or which part of the world they live in. can be fine-tuned. .
AI-powered generative chat
CEO Mark Zuckerberg said the company is focused on creating “AI personas that can help people in many ways.” Perhaps this ties in with the company’s plans to incorporate generative AI into its chat technology. This will allow you to talk to these characters via the company’s chat his platform (the biggest being his Whatsapp and Messenger) in order to interact with Meta’s various services. It also allows businesses to implement these services on their Facebook page and Whatsapp channel, allowing any business to effectively implement their own AI-powered automated customer service and feedback his agents. will be available to
image generation
Meta’s Facebook AI division has developed a unique image generation technology, instance-conditional adversarial generative networks (IC-GAN). According to the researchers, unlike standard his GAN-based image generators, it can be used to create a greater variety of images than those contained in the training dataset. It is predicted that one of the most useful uses of generative AI will be in the creation of synthetic data for training other machine learning algorithms, which is the use of richer synthetic training data from smaller sets. means that it is possible to create a set of Real-world training data. This can potentially reduce the cost of generating, collecting, and storing data for training AI algorithms. There is also a text-to-video generative AI application called . make a videosaid it plans to incorporate it into Reels’ short-form video platform in the future.
Natural language generation
Language-based generative AI applications such as the aforementioned chat functionality may eventually be powered by LLaMA – Large Language Model Meta AI –. Meta’s own answer To ChatGPT and Google’s Bard.
LLaMA is intentionally designed as a small language model, with its largest model being trained with 65 billion parameters, as opposed to the reported 1 trillion parameters for GPT-4. The benefit is that less computational power and resources are required for retraining to test new approaches and use cases. Smaller models with up to 7 billion parameters are also available. Such a model could potentially run on much smaller devices than the cloud servers required for ChatGPT or Bard, paving the way for self-contained instances running on personal computers and smartphones. . This can have important implications for companies that want to use generative language models while keeping their data private.
Generative AI and the Metaverse
In late 2021, the company, formerly known as Facebook, rebranded itself to Meta, declaring that its future lies in the Metaverse. While there is much debate about the exact meaning of the term, it is usually associated with the Internet, which is characterized by more immersive environments that may be rendered in virtual reality (VR), avatars, and shared online experiences. refers to the “next generation” iteration of
Since then, Meta’s stock has plummeted, job cuts have taken a toll, and revenues across the advertising platform have plummeted. Some commentators attribute this problem, at least in part, to the company’s, and particularly Zuckerberg’s, focus on making the leap into the metaverse, a concept that hasn’t yet been enthusiastically embraced by the general public.
But despite switching focus to AI in recent months, Mehta and Zuckerberg are still sticking to their company policy to some extent. They argue that the Metaverse will be a key component of his AI vision.
Meta’s own metaverse platform, Horizons, is built around creativity and specifically designed to allow users to build their own homes and environments within a VR environment. The company strongly suggests that its generative AI technology shows its true value here. CTO Andrew Bosworth said: “In the future, we may be able to simply describe the world we want to create and have a large language model generate that world.
“So more people have access to things like content creation.”
This allows the company to make it much easier for users to join and start creating virtual worlds without having to learn how to use a complex interface to build and deploy their dream metaverse home. It could mean that you want to promote the spread of your environment. 3D structure. Perhaps they hope that this democratizing effect will help the company’s billion-plus customer base make the leap from Facebook’s two-dimensional page to the three-dimensional world of Horizons. .
