Metalearning: Sparse Variable-Structure Automata
This work addresses a specific bottleneck in autoencoder design for machine learning practitioners, offering an incremental improvement in dynamic sparse coding.
The paper tackles the problem of selecting the optimal encoder dimension in autoencoders to balance reconstruction accuracy and computational complexity, proposing a metalearning approach that dynamically adjusts the number of basis vectors to meet error thresholds, achieving control over both dimension and error in an online framework.
Dimension of the encoder output (i.e., the code layer) in an autoencoder is a key hyper-parameter for representing the input data in a proper space. This dimension must be carefully selected in order to guarantee the desired reconstruction accuracy. Although overcomplete representation can address this dimension issue, the computational complexity will increase with dimension. Inspired by non-parametric methods, here, we propose a metalearning approach to increase the number of basis vectors used in dynamic sparse coding on the fly. An actor-critic algorithm is deployed to automatically choose an appropriate dimension for feature vectors regarding the required level of accuracy. The proposed method benefits from online dictionary learning and fast iterative shrinkage-thresholding algorithm (FISTA) as the optimizer in the inference phase. It aims at choosing the minimum number of bases for the overcomplete representation regarding the reconstruction error threshold. This method allows for online controlling of both the representation dimension and the reconstruction error in a dynamic framework.