There are a lot of new really exciting machine learning approaches that are emerging every day. The ones that we use in the Palop lab at the Gladstone Institutes in the study of animal models of disease are of two varieties. We use supervised learning tools to track the animal as it moves around the chamber, and we use an unsupervised learning tool in order to find the patterns in the animal’s motion...
There are a lot of new really exciting machine learning approaches that are emerging every day. The ones that we use in the Palop lab at the Gladstone Institutes in the study of animal models of disease are of two varieties. We use supervised learning tools to track the animal as it moves around the chamber, and we use an unsupervised learning tool in order to find the patterns in the animal’s motion. This is really essential because you can’t make detailed inferences around the disease phenotype from raw data. We must create methods to identify where in this incredibly rich data the disease phenotype can be identified so that we can find the biomarkers that can be targeted through therapeutic interventions. How we’re going to be able to translate these methods into the clinic. I think that for now, we are really working hard to make sure they are validated in preclinical and translational research, but the implications for patient care are clear. I’m particularly excited about how these approaches can be used to improve the sensitivity of diagnostics. And so if we’re able to identify disease much earlier, then we can hopefully intervene in the course of disease progression and stop people from getting sick.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
