Definition of AI Technology Development Tree
Artificial Intelligence Industry Analytics, a subsidiary of the Deep Knowledge Group developed an AI Technology Development Tree, system that collect existing models in the field of artificial intelligence and distributes them according to the algorithms that are used in these models.
This distribution has a tree structure, where the root is "Artificial Intelligence", closer to the root are more general categories of algorithms (e.g., reinforcement and supervised/unsupervised learning), at the end of the branch are single models (GPT-3, BERT, XGBoost etc.). For each category from penultimate layer we take only most popular models or models that contributed the most to the progress of this direction.
AI encompasses a broad range of algorithms and techniques. Classifying them helps in organizing and categorizing these algorithms based on their characteristics, underlying principles, and approaches. It provides a structured framework for understanding the different types of AI algorithms and their functionalities.
Please note, that this tree is about algorithms, not about tasks, that can be solved with help of AI! As an example, large language models, like GPT-4, can solve various natural language processing tasks (text generation, text classification, summarization etc.) and task of generating images can be solved by many algorithms, like GANs and diffusion models.