Preventing Catastrophic Forgetting in an MLP Online Learning Setting
To prevent catastrophic forgetting on seasonal weather data in a deep network, we designed a hybrid regression/auto-encoder network and found some improvemen...
To prevent catastrophic forgetting on seasonal weather data in a deep network, we designed a hybrid regression/auto-encoder network and found some improvemen...
Combining an MLP and LSTM, we compute the similarity between a given sentence and all images in a database to recommend the best image to represent the text.
To improve image classification when using transfer learning from dataset A to B, we tried to increase feature extraction in the pre-trained dataset A.