The era of enterprise AI has crashed down on us in recent months. The public infatuation with ChatGPT since its launch last November opened the floodgates of corporate interest and triggered an industry-wide land grab with all the major tech entities vying to claim their right in this burgeoning market by incorporating generative AI features. on your existing products. Heavyweights including Google, Microsoft, Meta and Baidu are already racing their LLMs for market dominance, while everyone else from Adobe and AT&T to BMW and BYD struggle to find uses. for revolutionary technology.
NVIDIA’s latest cloud service offering, AI Foundations, will enable companies strapped for time and money to develop their own models from scratch to “build, refine, and operate custom big language models and generative AI models that are They train with their own proprietary data created for their unique domain-specific tasks.”
These models include NeMo, NVIDIA’s text-to-image engine and competitor to DALL-E 2; BioNemo, a drug and molecule discovery-focused fork of the NeMo model built for the medical research community; and Picasso, an AI capable of generating images, videos and “3D applications… to increase productivity for creativity, design and digital simulation,” according to Tuesday’s statement. Both NeMo flavors are still in early access and Picasso remains in private preview despite Tuesday’s news, so it will be a minute before either releases to the general public. Both NeMo and Picasso run on NVIDIA’s new DGX Cloud platform and will eventually be accessible through an online portal.
These enterprise-facing cloud-based services function as blank templates that companies can pour their own databases into for specific training. So while something like Google’s Bard AI trains on (and will pull) data from all over the internet to provide a generated response, NVIDIA’s AIs will allow companies to tailor a similarly styled LLM to their own specific needs using their own proprietary data. — think ChatGPT but only for the research division of a pharmaceutical company. Models can be trained with between 8 billion and 530 billion parameters, more than triple the 185 billion parameters provided by GPT-3.5.
Imagine StableDiffusion, but trained on Getty Images with actual permission from Getty. NVIDIA announced one such system on Tuesday based on the NeMo cloud service: a series of responsibly sourced text-to-image and text-to-video models, “trained on the fully licensed assets of Getty Images,” were read in Tuesday’s press release. “fake images will provide royalties to artists on the revenue generated by the models.”
BioNeMo uses the same technical underpinnings as NeMo itself, but is fully geared towards drug and molecule discovery. According to Tuesday’s statement, Bio NeMo, “enables researchers to tune generative AI applications on their own proprietary data and run AI model inference directly in a web browser or through new cloud APIs that are easily integrated.” in existing applications.
“BioNeMo is dramatically accelerating our approach to biologics discovery,” Peter Grandsard, executive director of Biologics Therapeutic Discovery at Amgen, said in a statement. “With it, we can pre-train large language models for molecular biology on Amgen’s proprietary data, allowing us to explore and develop therapeutic proteins for the next generation of drugs that will help patients.”
Six models will be available at launch, including DeepMind’s AlphaFold2, Meta AI’s ESM2, and ESMFold ProtGPT-2, DiffDock, and MoFlow predictive models. According to the companies, the addition of AI-based predictive models helped reduce the time to train “five custom models for molecule detection and optimization” using Amgen’s proprietary data on antibodies from the usual three months to four. weeks.
NVIDIA also announced a similar partnership with Shutterstock. The photography site will use Picasso to generate 3D objects from text messages as a new feature within Creative Flow, with plans to offer it on Turbosquid.com and NVIDIA’s upcoming Omniverse platform.
“Our Generative 3D partnership with NVIDIA will power the next generation of 3D contribution tools, greatly reducing the time it takes to create beautifully textured and structured 3D models,” Shutterstock CEO Paul Hennessy said in the statement. “This first-of-its-kind partnership furthers our strategy of leveraging Shutterstock’s massive set of metadata to bring new products, tools and content to market. By combining our 3D content with NVIDIA’s core models and utilizing our respective marketing and distribution platforms, we are able to capitalize on an extraordinarily large market opportunity.”
NVIDIA is also partnering with Adobe as part of the latter’s Content Authenticity Initiative, which seeks to improve transparency and accountability within the AI generative training process. The CAI proposals include a “do not train” list, similar to robot.txt but for images and multimodal content, and persistent source tags that will detail if and from where a piece is AI-generated. The two companies have also announced plans to incorporate many of Picasso’s features directly into Adobe’s suite of editing software, including Photoshop, Premiere Pro and After Effects.
