At the ISC23 keynote, Intel announced Aurora genAI – a science-focused generative AI model with a trillion parameters, almost six times more than in the free and public versions of ChatGPT. This news has sparked conversations about all the possibilities this model can unlock.
It has always been well understood that to train and build models near human standards, companies require a humongous amount of computational power, where fine-tuning begins at the hardware level.
Intel’s bold vision makes a solid case backed up by one of the largest chip manufacturers. It has already proved its capability to produce a chip that matches and is often treated as a gold standard in compatible hardware for AI upscaling.
Backed by 2 Exaflops Intel’s Aurora Supercomputer, with Megatron and DeepSpeed models as its foundation, the Aurora-GenAI model promises to train scientific data, general data, scientific and machine codes, and other texts related primarily to the scientific domain with 1 trillion parameters which are almost six times the parameters we see in open and publically accessible versions of ChatGPT.
Intel is focussing on building this model to cater to the science community and accelerate the advancement in System Biology, Cancer Research, Climate Science, Cosmology, Polymer Chemistry, and Materials Science.
The Deep learning models we use today are well-trained in solving systematic problems. These systems can translate anything you can put down in a step-by-step manner. You can upscale it and use it on the fly to solve real-world problems. Besides the obvious existing use cases, people now expect to find patterns among complex data, such as molecular biology and formulation. Things like molecular binding patterns and compatibility revelations in a way that takes a lot of work to wrap one’s head around with conventional methods.
What is more interesting is that Intel aims to predict the patterns and bottlenecks we leave out due to a lack of vision and understanding of the use case at hand, especially when it is coupled with time. To understand it, This model will aim to predict the problematic scenarios we can’t estimate or see yet but are likely to come up at some point in time, given what the data is about that particular problem.
People are taking this announcement with lots of excitement and positivity in the AI community. People are more invested in knowing how it would perceive and understand the topics that are, by nature, more sensitive and challenging, for example, political scenarios and policy-making, prevalent social issues, climate changes, cosmological predictions, and its take on it to solve them to some level.
It is interesting to understand here that this project is a work in progress right now, and it is talking about a future commitment. In reality, it is still, in fact, an announcement. The project will be developed in collaboration with Argonne National Laboratory and HPE.
In conclusion, this news brings a lot of hope not only to the AI community but also to retail investors. This news drives positive sentiments up for Intel, making it a promising stock option to explore, which certainly puts Intel in a good spot. It would be interesting to see how Intel will play out against some of its closest competitors in the market, such as Nvidia, and how well its model will adapt to the commitments made.
Check Out The Announcement. Don’t forget to join our 22k+ ML SubReddit, Discord Channel, and Email Newsletter, where we share the latest AI research news, cool AI projects, and more. If you have any questions regarding the above article or if we missed anything, feel free to email us at Asif@marktechpost.com
Anant is a Computer science engineer currently working as a data scientist with experience in Finance and AI products as a service. He is keen to build AI-powered solutions that create better data points and solve daily life problems in an impactful and efficient way.