Artificial intelligence (AI) doesn’t work well in situations where little data is available. Drug development is one of this scenarios, but researchers at Stanford University believe a relatively new type of deep learning might be the answer to AI’s low-data problem.
Vijay Pande, professor of chemistry at Stanford University, and his students are analyzing how one-shot learning might work when it comes to the early stages of drug development. Unlike other forms of machine learning, one-shot learning can learn with access to only a small number of data points.
Pande’s group initially thought using one-shot learning for early stage drug design was farfetched, but the results published April 3 in ACS Central Science show one-shot learning has potential for drug development and other areas of chemistry research.
“We’re trying to use machine learning, especially deep learning, for the early stage of drug design,” says Pande. “The issue is, once you have thousands of examples in drug design, you probably already have a successful drug.”
Deep Learning: From Identifying Images to Molecules
One-shot learning has been used successfully for image recognition and genomics, but applying it to problems relevant to drug development is a bit different. Whereas pixels and bases are fairly natural types of data to feed into an algorithm, properties of small molecules aren’t.
To make molecular information more digestible, the researchers first represented each molecule in terms of the connections between atoms. This step highlighted intrinsic properties of the chemical in a form that an algorithm could process.
Stanford chemistry Professor Vijay Pande and his students see a future for one-shot learning in the early stages of drug development. (Image credit: L.A. Cicero)
With these graphical representations, the group trained an algorithm on two different datasets:
- One with information about the toxicity of different chemicals
- One that detailed side effects of approved medicines
From the first dataset, they trained the algorithm on six chemicals and had it make predictions about the toxicity of the other three. Using the second dataset, they trained it to associate drugs with side effects in 21 tasks, testing it on six more.
In both cases, the algorithm was better able to predict toxicity or side effects than would have been possible by chance.
“We worked on some prototype algorithms and found that, given a few data points, they were able to make predictions that were pretty accurate,” says Bharath Ramsundar, who is a graduate student in the Pande lab and co-lead author of the study.
However, Ramsundar cautioned that this isn’t a “magical” technique. It was built off of several recent advances in a particular style of one-shot learning and it works by relying on the closeness of different molecules, as indirectly indicated by their formula. For example, when the researchers trained their algorithm on the toxicity data and tested it on the side effect data, the algorithm completely collapsed.
One-Shot Learning Won’t Take Jobs
People concerned about AI taking jobs from humans have nothing to fear from this work. The researchers envision this as groundwork for a potential tool for chemists who are early in their research and trying to choose which molecule to pursue from a set of promising candidates.
“Right now, people make this kind of choice by hunch,” Ramsundar says. “This might be a nice compliment to that: an experimentalist’s helper””
Beyond giving insight into drug design, this tool would be broadly applicable to molecular chemistry. Already, the Pande lab is testing these methods on different chemical compositions for solar cells. They have also made all of the code they used for the experiment open source, available as part of the DeepChem library.
“This paper is the first time that one-shot has been applied to this space and it’s exciting to see the field of machine learning move so quickly,” Pande said. “This is not the end of this journey – it’s the beginning.”