Learning to Learn, Meta Learning, machine learning deep learning
This project should be compete with the state-of-the-art similar projects. Say, model augmentation via external memory as in One-shot Learning with Memory-Augmented Neural Networks is basically reading and writing from a matrix dummy memory. An objective is glaring here. When replacing the memory by a DNN, retrieving them verbatim. The goals of the MAML techniques in the field of modeling. Obviously, it wouldn't be an effective resource when it comes to the learning process.
The system structure appears to be very much detailed. The system design is built-up well-defined functionaries. This makes replacing or augmenting new component an easy task but calculatable. Over more, the system in Learning Unsupervised Learning Objective Rules was Intended for unsupervised learning, Which opens the Possibility to tweak it in other tasks Such as reinforcement learning.
MNIST, MNIST-1K, CelebA, Omniglot, CIFAR-10, glyph, CIFAR-100 and ImageNet. Those datasets are image-based datasets and they are accessible and downloadable.
gépi tanulás, machine learning, deep learning, GAN