Personalization with Meta Learning

Expansion of the volume of available information has made it one of the most important issues today to access the information that each person wants in the fastest possible time. One of the solutions to this problem is the idea of personalization. Personalization in general means turning the available information and user interface into the best and most ideal state for each particular user. The application of this issue can be in various fields such as electronic services, social networks, medicine and health, etc. So far, various solutions to the problem of personalization have been proposed. One of the methods that has received more attention in recent years is personalization using meta-learning algorithms. These algorithms can be widely used in this field due to the continuous improvement of the learning process and the ability to learn using small data. Paying attention to the challenges of a personalization issue also helps us to provide a comprehensive solution. Challenges such as information privacy, cold start, sequential learning, interpretability, etc. are among the challenges that exist in machine learning issues and specifically in personalization issues. In this project, we try to provide a personalization framework based on meta learning and also considering solving existing challenges.