SYNB1 - Machine Learning for Protein Design

Part of Canadian Synthetic Biology Research Group (CSBERG)’s summer workshop on synthetic biology. Timestamps: 9:30 - Activation functions 11:52 - Geometry of gradient descent 16:47 - Mean squared error for regression 19:10 - Cross entropy loss and KL divergence 28:31 - Ensembles 30:26 - Model performance 33:00 - Transfer learning 36:37 - Interpretability * (actually it’s neurons on the cortex ) 41:27 - Protein representations *