Fig. 3 | Scientific Reports

Fig. 3

From: A meta-learning framework to mitigate negative transfer in transfer learning applicable to drug design

Fig. 3

Model architectures. Shown is a schematic representation of the base model and meta-model. The base model is a feedforward neural network that takes a compound fingerprint as input and predicts the class label of this compound. The meta-model is a modular neural network that takes a compound fingerprint and a one-hot encoded representation of the target protein as input and predicts a weight for this compound-protein pair.

Back to article page