Intel has announced Aurora genAI, a new generative AI model for science. The 2 Exaflops Aurora supercomputers will be used to power the Aurora genAI artificial intelligence model. Announced in the ISC23 keynote, the Aurora genAI model will be trained on general text, scientific texts, scientific data and field-related codes. This model will be a fully science-driven generative AI model with potential applications.Areas of focus for Aurora genAI
- Systems Biology
- Cancer Research
- Climate Science
- Cosmology
- Polymer Chemistry and Materials
- Science
This AI model will have applications in many applications, from the design of molecules and materials to the synthesis of information from millions of sources, to proposing new experiments in systems biology, polymer chemistry and energy materials, climate science and cosmology. The model will also be used to accelerate the identification of biological processes associated with cancer and other diseases and to suggest targets for drug design.
1 trillion parameters target
The Intel Aurora genAI model is based on Megatron and DeepSpeed. Most importantly, the targeted size for the new model will be 1 trillion parameters. By the way, let’s also mention that the target size for the free and public versions of ChatGPT is only 175 million. This means a 5.7 times increase in the number of parameters.
41200
