الفهرس | Only 14 pages are availabe for public view |
Abstract A good machine learning model often requires training with a large number of samples. Humans, in contrast, can generalize, learn new concepts and skills much faster and more efficiently. Kids who have seen cats and birds only a few times can quickly tell them apart. They conclude knowledge from their prior experiences, as they are able to combine previous observations with small amounts of new evidence to learn fast. Is it possible to design a machine learning model that is able to learn new concepts and skills fast with a few training examples? That’s essentially what meta learning aims to solve. Meta learning, or learning to learn, gained a huge rise in interest in recent years. Contrary to traditional approaches to artificial intelligence where a given problem is solved from scratch using a standard learning algorithm. what meta learning aims at is to improve the learning algorithm itself. In this thesis, we will discuss about artificial intelligence techniques, machine learning, deep learning, deep generative models, the recent advances and approaches to meta learning, meta learning problem formulation, and finally propose an approach to enhance a type of meta learning algorithms called gradient-based algorithms. |