Episode 43 of The Data Science Podcast with Fexingo. Lucas and Luna explore how machine learning is transforming medical diagnosis — specifically, the case of a deep learning model developed at Stanford that detects skin cancer with accuracy comparable to board-certified dermatologists. Published in Nature in 2017, the model was trained on nearly 130,000 images representing over 2,000 diseases. Lucas breaks down how the team used a GoogleNet Inception v3 architecture, fine-tuned on a dataset that included both clinical and dermoscopic images. They discuss the challenges of dataset bias (the training data was predominantly light-skinned), the regulatory hurdles for deploying such models in clinics, and the current state of FDA-approved AI diagnostic tools as of mid-2026. Luna asks about the reproducibility crisis and how to trust a model that can't explain its reasoning. A concrete look at one of the most promising — and fraught — applications of data science today.

#MachineLearning #MedicalDiagnosis #DeepLearning #Stanford #AIinMedicine #SkinCancerDetection #GoogleNet #Nature #HealthcareAI #FDA #AlgorithmicBias #Explainability #DataScience #Technology #FexingoTechnologyShow #FexingoBusiness #BusinessPodcast #Fexingo

Keep every episode free: buymeacoffee.com/fexingo