Overcomplete autoencoder           Under-complete autoencoder           Sparse autoencoder           Complete autoencode

It seems like you're asking for explanations of different types of autoencoders in machine learning. Here they are:

1. Overcomplete Autoencoder: This type of autoencoder has a higher dimensionality in the hidden layer compared to the input layer. This might lead to a simple identity function as a mapping function, which is not useful. To avoid this, we can add constraints such as sparsity or denoising.

2. Under-complete Autoencoder: In this type, the dimensionality of the hidden layer is less than the input layer. This forces the autoencoder to learn the most important attributes of the input data, thus achieving data compression.

3. Sparse Autoencoder: This type of autoencoder adds a sparsity constraint on the hidden layer, which forces only a small number of neurons to activate at a given time. This leads to a more robust model that captures high-level features in the data.

4. Complete Autoencoder: This term is not standard in the field of machine learning. It might refer to an autoencoder where the dimensionality of the hidden layer is the same as the input layer. However, without additional constraints, this type of autoencoder might not be useful as it could simply learn the identity function.