Professor Frederick Sanger, a British biochemist, received his second Nobel Prize in Chemistry in … (+)
It seems that the average person is quite familiar with generative AI. These polished, image-generating consumer products like Dall-E are, to some extent, household names. But fewer of us realize how far generative AI can go to introduce new solutions to our world.
This is particularly true at the biological level. In the same way that genAI can “hack” the ability to simulate text, images, and video, it can look to biology and find new ways to identify biological structures.
Consider this example: A Microsoft team is figuring out how to use a diffusion model to find new materials that can meet the needs of high-tech projects.
“Finding a new material for a target application is like looking for a needle in a haystack,” write the authors of a blog article at Microsoftwhere they worked on just such a program, something called, aptly, MatterGen. “Historically, this task has been accomplished through costly and time-consuming experimental trial and error. More recently, computational screening of large databases of materials has allowed researchers to accelerate this process. However, finding the few materials with the desired properties still requires screening millions of candidates.
In a way, this reminds us why AI is the ideal partner for the Internet. You can take hours, days, or even weeks to search as much as possible on the internet for a particular fact or piece of information – or you can ask a model like ChatGPT and get it in seconds. The Internet is a vast trove of information that stretches across the globe – AI is an ultra-productive researcher capable of aggregating results from the Internet in the blink of an eye.
It’s simply orders of magnitude more efficient than a human at searching the Internet. It is also apparently several orders of magnitude more efficient than a human at finding new types of matter, materials with unique chemistry, magnetism, conductivity, etc. specific.
Diffusion architecture and the biological world
To understand what MatterGen does with biological science research, you need to understand the diffusion model itself.
In general, when generative AI is applied to broadcasting, you start from a defined object, an image or a protein or some kind of biological element, for example, and you corrupt it, with a “foley” process. . You introduce noise, to diffuse the original element into an abstract. Then the system turns around and “denoises” the target, merging it into a new result, element, or structure that has the attributes or properties desired by the researcher, the human user.
Microsoft researchers report that when the system received more than 608,000 stable materials from databases, it was able to generate good candidates for new materials.
Coping with Compositional Disorder
One challenge, the team noted, is compositional disorder, in which atoms can change position within a particular synthesized material.
“The community has been exploring what it means for a material to be novel in the context of computer-designed materials, as widely used algorithms will not distinguish between pairs of structures where the only difference is a permutation of elements similar in their respective sites.” explains the team.
Clear as mud?
As I mentioned before, you can also use ChatGPT to better understand what these researchers are talking about. I put the above into ChatGPT and got this.
“Recently, researchers have looked at what makes a material “new” when it comes to computer-designed materials. This is important because common algorithms often cannot differentiate between structures that differ only in how similar elements are arranged in their positions.
I also asked for a concrete example of compositional disorder, and got this:
“A concrete example of compositional disorder can be found in metal alloys, such as brass. Brass is made by combining copper (Cu) and zinc (Zn). In a perfectly ordered structure, the copper and zinc atoms would be arranged in a regular pattern. However, in reality, these atoms can change places randomly in the crystal lattice, meaning that some sites thought to contain copper atoms could contain zinc atoms, and vice versa. This randomness in the arrangement of copper and zinc atoms is an example of compositional disorder, which can influence material properties, such as strength, corrosion resistance, and electrical conductivity.
I learned a lot about the attributes of brass and why it is soft.
“Brass is generally considered soft, especially when it contains a high percentage of zinc,” ChatGPT provided. “Although compositional disorder can influence material properties, brass’s softness is primarily due to the way copper and zinc atoms interact in its crystal structure. When making brass, zinc atoms can replace copper atoms in the crystal lattice. This creates a solid solution, in which the different sizes and properties of the copper and zinc atoms disrupt the ordered arrangement of the atoms, making it easier for layers of atoms to slide past each other. This sliding is what makes the material ductile (can be stretched into wires) and malleable (can be hammered into shapes), characteristics of softness. So while compositional disorder might contribute to the softness by making the structure less stable, the main reason for brass’s softness is the specific arrangement and interaction between the copper and zinc atoms. The relative proportions of these metals in the alloy also play a key role; higher zinc content generally leads to softer brass.
One of the key points here is that some alloys can have combinations of smaller and larger atoms, with the smaller atoms fitting between the larger ones. It is the compositional parity of the copper and zinc atoms that leads to some of the attributes of the brass material. The model and other sources also point out that when you add more zinc, the brass becomes softer.
Innovative battery technology
As soon as this new generative AI solution is announced, the Microsoft team talks about innovative lithium-ion batteries and how the original lithium-ion battery was designed with the discovery of lithium-cobalt oxide in the 1980s. As they mention, lithium-ion batteries are essential for everything from modern smartphones to electric and hybrid vehicles; to do without it would set our world back considerably.
Now, researchers are trying to figure out how to replace traditional lithium-ion batteries with better, more modern battery designs — ones that use less lithium, because it’s difficult to obtain and extract.
And they succeed in using this biological technology. Going back about a year you can see results from PNNL (The Pacific Northwest National Laboratory), where researchers obtained a design requiring 70% less lithium.
It is likely that we will continue to achieve better iterations through the use of the AI process, where diffusion shows us what may be possible in chemical engineering.
I’ve written about other AI-based biological solutions before, but this one is exciting in part because of the potential for advancements in lithium-ion batteries and other critical changes in how we source our energy. business materials.
Think about it in simple logistics terms: you want a better supply chain, safer materials, higher quality projects, and better delivery to customers – and all of this can help us get there.