Introduction One thing newcomers to machine learning (ML) and many experienced practitioners often don’t realize is that ML doesn’t extrapolate. After training an ML model on compounds with µM potency, people frequently ask why none of the molecules they designed were predicted to have nM potency. If you’re new to drug discovery, 1nM = 0.001µM. A lower potency value is usually better. It’s important to remember that a model can only predict values within the range of the training set. If we’ve trained a model on compounds with IC50s between 5 and 100 µM, the model won’t be able to predict an IC50 of 0.1 µM. I’d like to illustrate this with a simple example. As always, all the code that accompanies this post is available on GitHub.

A combined team of roboticists from CREATE Lab, EPFL and Nestlé Research Lausanne, both in Switzerland, has developed a soft robot that was designed to mimic human infant motor development and the way infants feed.

In their paper published in the journal npj Robotics, the group describes how they used a variety of techniques to give their robot the ability to simulate the way human infants feed, from birth until approximately six months old.

Prior research has shown that it is difficult to develop invasive medical procedures for infants and babies due to the lack of usable test subjects. Methods currently in use, such as simulations, observational instruments and imaging tend to fall short due to their differences compared to real human infants. To overcome such problems, the team in Switzerland has designed, built, and tested a soft robotic infant that can be used for such purposes.