IBM is launching a new open-source foundational artificial intelligence (AI) model for chemistry that aims to accelerate the discovery of new, more sustainable materials with applications in chip manufacturing, clean energy, and consumer packaging. announced the family. According to the company, AI has the potential to transform the discovery of safer and more sustainable materials, offering a powerful solution for phasing out hazardous substances tracked by the U.S. Environmental Protection Agency. I will.
Also read: IBM releases new AI models built for business
Material discovery using AI
IBM Research designed these new AI models to accelerate this process by screening millions of molecules for desirable properties and generating new, safe alternatives. IBM says the model, available on platforms such as GitHub and Hugging Face, has been downloaded more than 100,000 times, reflecting strong interest in its potential.
“Using foundational models pre-trained on vast molecular databases, we can screen millions of molecules at once for desirable properties while weeding out molecules with dangerous side effects. These models “It is also possible to generate entirely new molecules in nature using traditional, long-term, trial-and-error-based discovery processes,” IBM said on Dec. 20.
Also read: Meta launches new AI models and tools to drive innovation
overcome challenges
The main challenge in applying AI to chemistry lies in effectively representing molecular structures. IBM researchers have addressed this problem by integrating various data formats such as SMILES, SELFIES, and molecular graphs into a unified “multi-view” Mix of Experts (MoE) model. IBM says the adaptive model improves task-specific performance and outperforms existing models in predicting properties such as toxicity and solubility, which are important benchmarks in drug and materials discovery.
IBM also announced plans to release new fusion technologies and models next year built on additional data modalities, such as positioning atoms in 3D space. Through the AI Alliance, the company also collaborates with other academic and industry researchers to accelerate the discovery of safer and more sustainable materials.
Collaboration for global sustainability
Earlier this year, IBM and Japanese materials company JSR established the Working Group on Materials (WG4M), which includes approximately 20 corporate and academic partners. The group focuses on developing foundational models, datasets, and benchmarks to address challenges ranging from creating reusable plastics to advancing materials for renewable energy.
Also read: IBM study predicts vehicles will be software-defined and AI-powered by 2035
The power of AI
Through continued collaboration with industry and academia, IBM is committed to driving the development of new foundational models and datasets to address global challenges, such as creating reusable plastics and materials that support clean energy. That’s what I’m aiming for. IBM emphasizes that AI has the power to increase creativity and encourage the search for safer and greener alternatives across industries.
Dave Brains, CTO of Emerging Technologies at IBM Research UK, said: “From semiconductor manufacturing to clean energy, new, more sustainable materials are needed in almost every industry. It gives me the power to increase my creativity many times over.”
Subscribe
