1987-1993 (AI Winter 2): Collapse of the expert systems market leads to another funding drought
1997: IBM’s Deep Blue defeats world chess champion Garry Kasparov
2000s: Machine learning and statistical methods dominate AI; rise of support vector machines, Bayesian networks
2011: IBM’s Watson wins Jeopardy! against human champions
2012: Deep learning breakthrough: Alexnet (by Hinton, Krizhevsky, Sutskever) wins the ImageNet competition, sparking the modern AI revolution
2016: AlphaGo defeats world champion Lee Sedol in the game of Go - a milestone for deep learning
2017-2018: Transformer Architecture is introduced in the paper Attention Is All You Need, revolutionizing NLP. Large Language Models (LLMs) like BERT & GPT-2 show general purpose NLP capabilities
2020: GPT-3 (OpenAI) demonstrates seemingly human-like text generation
2022-2023: Generative AI boom: Stable Diffusion & DALLE-2 bring text-to-image capabilities. ChatGPT takes off
2024-2025: Multimodal AI (GPT-4, Gemini, Claude) emerge: capable of understanding text, images, audio, etc. Current AI research focuses on scaling up these systems