This technique is called supervised learning because the AI developers used a dataset containing example inputs (in this case, molecular structures) that were individually labeled according to the desired output or result (in this case, effectiveness as an antibiotic).
