Artificial intelligence as a tool for science

AI systems have already found their way into science, technology and medicine in recent years. This trend is continuing and intensifying. This year, for example, there are also approvals for AI systems in radiology, a dermatology app from Google, an AI improvement for microscopy or an AI system for chip design. Other AIs help develop drugs or find better materials.

As early as December 2020, Deepmind will show AlphaFold 2.0, an AI system considered a scientific breakthrough because it can solve the age-old problem of protein folding, advancing many areas of research.

This year, the first effects are starting to show: AlphaFold is being used, for example, by the Geneva Initiative for Neglected Diseases to cure parasitic diseases. Other teams are using AlphaFold to calculate the spike protein structure of new COVID variants, for example.

Then, in July, Deepmind releases AlphaFold 2 as open source, adding 350,000 3D structures from 20 different organisms to the open-access “Protein Data Bank.” About 20,000 of the predicted structures belong to humans, representing about 98 percent of the protein structures in the human body. Science names AlphaFold “Breakthrough of the Year.”

AI in modern medicine

Medicine is on the verge of another revolution: Artificial intelligence supports doctors in analyzing X-ray and ultrasound images as well as in diagnostics and treatment. Researchers in Germany are also working successfully on novel solutions – and good access to data is important here.

Artificial intelligence (AI) is an important driver of the digital revolution. Health research and care are already important fields of application. AI is particularly advanced in the analysis of medical images, for example. Here, the field of “computational photonics,” the combination of modern photonic processes with fast and intelligent data analysis, promises significant innovations for medicine. The potential applications of AI are manifold.

Such innovations also hold great potential for companies in the healthcare industry. Through its artificial intelligence strategy, the German government will play a key role in shaping these developments, focusing in particular on their transfer to healthcare and care.

In modern biomedicine, huge and complex data sets are generated. They are generated in particular in the fields of imaging and ‘omics’, the analysis of different components of cellular metabolism on a global scale. The analysis and integration of these datasets is crucial not only for research, but also for planned clinical use, e.g. in precision medicine. The vision of a truly data-driven medicine is therefore enabled by data science techniques, especially machine learning and artificial intelligence.

Diagnosing and treating diseases based on their molecular causes – that is the goal of molecular medicine. Thanks to new technologies such as genome analysis and CRISPR screens, we have a great deal of data on how cancer develops. Machine learning helps us understand this data and make it useful for individual patients. With epigenetics, we are looking back into the origin of a tumor – to the cell that started it all. And we are looking ahead: What potential do immune cells have for tumor defense and how can we support them?

The computer can learn “supervised

We speak of artificial intelligence when computer programs are capable of learning. And there are two main ways for this learning. One way can be called “supervised learning”: Put simply, researchers show the computer very many similar things, teaching it what is right or wrong, healthy or pathological, a certain cell formation or just not the desired cell formation, depending on the question. The idea behind this: When the computer has received enough input, it can make the distinction itself. In this way, it can not only take a lot of work off the doctors’ hands, it can also improve the quality of diagnosis and treatment.