Update Log
2021-07-24: Initial Upload
Auto-Encoder
- It is an unsupervised approach for learning a lower-dimensional feature representation from unlabelled training data.

- The encoder is trained usually with a decoder trying to recover the original data . L2 loss between the reconstructed and original input data is calculated to facilitate the training process.

- The encoder alone can initialise a supervised learning without decoder as a feature extractor.

Variational Auto-Encoder
- Why VAE?
- Samples are discrete and can not represent the whole distribution 100% accurate. Let’s take the moon as an example.

- Samples are discrete and can not represent the whole distribution 100% accurate. Let’s take the moon as an example.