What exactly is the scientist AI – just a fancy neural network?
This is not one neural network, but rather a set of computer programs that can aid scientists make novel discoveries. My group has already developed algorithms that can aid in individual tasks, such as weather forecasting, identifying the factors of global temperature raise or an attempt to discover causal compounds, such as the impact of vaccination policy on the transfer of the disease.
We are now building a wider “Foundation”, which is versatile enough to support many tasks. Scientists collect data from all types of instruments and we want our model to contain various types of data, numbers, text, images and movies. We have an early prototype, but we want our model to be more comprehensive, more wise and better trained before its release. This can happen within a few years.
What are you imagining?
AI can aid at virtually every stage of the scientific discovery process. When I say: “scientist”, I really mean AI scientific assistant. For example, the stage of literature research in the experiment usually requires huge data collection and effort. But now a vast language model can read and summarize thousands of books during one break for lunch. What AI is not good is to assess scientific validity. In this case, he cannot compete with an experienced researcher. While AI can aid generate hypotheses, designing experiments and data analysis, it still cannot carry out sophisticated experiments.
How far would you like to see this concept?
When I imagine this, AI scientists can leisurely down researchers from some zerczarnia, while allowing people to deal with original aspects of science. This is something that it is especially good at. We ensure that the goal is not to replace human scientists. I cannot imagine that I would not like to see how the machine replacing or disturbing human creativity.
Original story reprinted with consent from How much warehouseeditorly independent publication Simons Foundation whose mission is to raise public understanding of science by covering the development of research and trends in mathematics and physics and life sciences.