Cost‑Effective Training of Deep CNNs with Active Model Adaptation

Deep convolutional neural networks have achieved great success in various applications. However, training an effective DNN model for a specific task is rather challenging because it requires a prior k