Human versus machine: AI narrowly beats human scholars in scientific ability test

Advertisement

Human versus machine: AI narrowly beats human scholars in scientific ability test

PISCATAWAY, NJ — No other invention represents human ingenuity and intelligence like the computer. A marvel of modernity, countless works of science fiction have predicted an inevitable confrontation in the not too distant future: man versus machine. According to researchers at Rutgers University, machines already appear to have surpassed humanity in at least one scientific discipline.

Professor Vikas Nanda of Rutgers University has spent over two decades meticulously studying the intricate nature of proteins, the highly complex substances found in all living organisms. He has dedicated his professional life to looking at and understanding the unique patterns of amino acids that make up proteins and determining if they become hemoglobin, collagen, etc. In addition, Prof. Nanda is an expert in the mysterious step of self-assembly, where certain proteins clump together into even more complex fabrics.

So when the study authors ran an experiment pitting a human — someone with a deep, intuitive understanding of protein design and self-assembly — against the predictive capabilities of an AI computer program, Prof. Nanda was the perfect participant.

The study’s authors wanted to see who or what could be better at predicting which protein sequences would combine most successfully — Prof. Nanda and several other humans, or the computer. The published results show that the intellectual battle is near, but the AI ​​program has beaten the humans by a small margin.

What can scientists use the self-organization of proteins for?

Modern medicine is investing heavily in protein self-assembly, as many scientists believe that a full understanding of the process could lead to numerous revolutionary products for medical and industrial purposes, such as: B. artificial human tissue for wounds or catalysts for new chemical products.

“Despite our deep expertise, the AI ​​performed as well or better on multiple data sets, demonstrating the tremendous potential of machine learning to overcome human bias,” says Nanda, professor in the Department of Biochemistry and Molecular Biology at Rutgers Robert Wood Johnson Medical School, in a university edition.

Proteins are made up of large amounts of amino acids connected end-to-end. These amino acid chains fold into three-dimensional molecules with complex shapes. The exact form is important; The exact form of each protein, as well as the specific amino acids it contains, determine what it does. Some scientists, including Prof. Nanda, regularly engage in an activity called “protein design,” which involves creating sequences that produce new proteins.

Recently, Prof. Nanda and a team of researchers designed a synthetic protein capable of rapidly recognizing the dangerous nerve agent called VX. This protein can lead to the development of new biosensors and treatment methods.

For reasons still unknown to modern science, proteins self-assemble with other proteins to form superstructures that are important in biology. Sometimes it seems as if the proteins follow a design, such as when they self-assemble into a protective outer shell of a virus (capsid). In other cases, however, proteins self-assemble seemingly in response to something going wrong, eventually forming deadly biological structures linked to diseases from Alzheimer’s to sickle cell.

“Understanding protein self-assembly is fundamental to advances in many fields, including medicine and industry,” adds Prof. Nanda.

How did the AI ​​program fare?

During the test, Prof Nanda and five other colleagues were given a list of proteins and had to predict which ones were likely to self-assemble. The computer program made the same predictions, and then the researchers compared the human and machine responses.

The human participants made their predictions based on their previous experimental protein observations, such as patterns of electrical charges and levels of water aversion. People eventually predicted that 11 proteins would self-assemble. The computer program, meanwhile, selected nine proteins via an advanced machine learning system.

The human experts got it right on six of the 11 proteins they selected. The computer program achieved a higher accuracy percentage because six of the nine selected proteins were actually able to self-assemble.

Study authors explain that the human participants tended to “favor” certain amino acids over others, leading to incorrect predictions. The AI ​​program also identified some proteins that were not “obvious choices” for self-assembly, opening the door for further research. Prof Nanda admits he was once a doubter of machine learning for protein assembly studies, but he is now much more open to the technique.

“We are working to gain a fundamental understanding of the chemical nature of the interactions that lead to self-assembly, so I was concerned that using these programs would prevent important insights,” he concludes. “But what I’m starting to really understand is that machine learning is just another tool, like any other.”

The study is published in the journal natural chemistry.

You May Also Like