The Economist: Can AI change the way scientific research is done?

Artificial intelligence (AI) is shaping the field of science in unprecedented ways. **From accelerating the research process to generating new research hypotheses, the addition of AI brings huge potential to science. **

Earlier this year, Yann LeCun, one of the godfathers of modern AI, said: “By augmenting human intelligence, AI may spark a new renaissance, perhaps a new stage of the Enlightenment.”

Today, AI can already make some existing scientific processes faster and more efficient, such as discovering new antibiotics, new materials for batteries and solar panels, as well as predicting short-term weather, controlling nuclear fusion, and more. Demis Hassabis, CEO of Google DeepMind, compared AI to a telescope and believed that “AI may bring about a renaissance of new discoveries and become a multiplier of human wisdom.”

However, can AI do more by changing the way science itself works?

Literature-based discovery: AI leads the discovery of scientific knowledge

In fact, this shift has happened before.

With the advent of the scientific method in the 17th century, researchers began to trust experimental observations and the theories derived from them rather than the conventional wisdom of antiquity. The establishment of research laboratories in the late 19th century spurred innovation in fields ranging from chemistry to semiconductors to pharmaceuticals. These shifts not only increase scientific productivity, they also transform science itself, opening up new areas of research and discovery.

So how could AI achieve a similar transformation back then, not just in generating new results, but in new ways of generating new results?

**One promising approach is literature-based discovery (LBD). **

As an AI method, LBD aims to make new discoveries by analyzing scientific literature. As early as the 1980s, Dr. Don Swanson of the University of Chicago established the first LBD system in order to find new associations in the medical journal database MEDLINE. One of the early successes of this approach was in linking Raynaud’s disease, a circulatory disease, to blood viscosity, leading to the hypothesis that fish oil might be useful in treatment, a hypothesis that was later confirmed experimentally. However, the reach of LBD systems at that time was limited.

Nowadays, AI has made significant progress in natural language processing (NLP), and the amount of scientific literature has also increased significantly, making LBD methods even more powerful. For example, in 2019, researchers at Lawrence Berkeley National Laboratory in the United States used unsupervised learning techniques to analyze abstracts of materials science literature and convert them into mathematical representations called “word embeddings.” This approach allows AI systems to gain “chemical intuition” and suggest new materials that might have specific properties. After experimental verification, all top ten candidate materials showed excellent performance.

A recent paper published in Nature Human Behavior by University of Chicago sociologists Jamshid Sourati and James Evans extends this approach in a novel way. Researchers trained a system to consider both concepts and authors and achieved better results than before. Furthermore, they require the system to avoid mainstream research directions and identify “alien” hypotheses that are unlikely to be discovered under normal circumstances. This approach not only helps accelerate scientific discovery, but also reveals new “blind spots.”

Today, LBD systems can not only generate new research hypotheses but also identify potential partners and facilitate interdisciplinary collaboration. The application of this method is expanding to handle different types of documents such as tables, charts and figures, providing wider support to scientists.

Robotic Scientist: AI Leads the Laboratory Revolution

**Robotic scientists represent another exciting development beyond traditional laboratory automation. **They acquire background knowledge about a specific research area in the form of data, research papers, and patents, then generate hypotheses, perform experiments, evaluate results, and ultimately discover new scientific knowledge.

“Adam” at Aberystwyth University is a pioneer of robotic scientists. It has achieved the first independent discovery of new scientific knowledge. The experiment on the relationship between genes and enzymes in yeast metabolism is a typical case.

More sophisticated robotic scientists, like “Eve,” use machine learning to create “quantitative structure-activity relationships” (QSARs) — mathematical models that relate chemical structures to biological effects — as they plan and analyze experiments. Eve has already been used in drug discovery, successfully discovering that triclosan, an antimicrobial compound used in toothpaste, inhibits a key mechanism in the parasite that causes malaria.

At one time, the prospect of machines beating the best human players seemed decades away, but technology is advancing faster than expected. As robotic scientists become more and more capable, it will be possible to pit future robotic scientists against AI systems that can play chess.

Ross King, an AI researcher at the University of Cambridge who created Adam, said, “If AI can explore the entire hypothesis space, or even expand this space, then it may show that humans are only exploring a small part of the hypothesis space, perhaps due to their own scientific biases. of."

Robot scientists have transformed scientific research in a unique way by solving efficiency problems in the scientific field. The efficiency of scientific research is gradually decreasing and it is difficult to promote the development of the frontier of knowledge. Robot scientists can solve this problem through AI-driven systems, because machines can perform laboratory work faster, cheaper, and more accurately than humans, and can work around the clock . In addition, they can provide reproducible experimental results and alleviate the reproducibility crisis.

The potential and challenges of AI in science

**While AI has great potential in science, it also faces some challenges. **

In addition to better hardware and software and tighter integration between the two, there is also a need for greater interoperability between laboratory automation systems, as well as common standards that allow AI algorithms to exchange and interpret semantic information. Another obstacle is scientists’ lack of familiarity with AI-based tools. Additionally, some researchers worry that automation will threaten their jobs.

However, the impact of AI is now “far-reaching and pervasive,” said Dr. Yolanda Gil, a computer scientist at the University of Southern California. Many scientists are now “actively looking for AI partners.” Awareness of the potential of AI is growing, especially in the fields of materials science and drug discovery, where practitioners are building their own AI systems.

Overall, scientific journals have changed the way scientists discover information and learn from each other. Research laboratories have expanded the scale of experiments and realized the industrialization of experiments. By extending and combining the first two revolutions, AI can indeed change the way scientific research is done.

Reference links: