terryum/awesome-deep-learning-papers

This is a curated reading list of the most influential deep learning research papers published between 2012 and 2016, the period when modern AI as we know it was essentially born. Deep learning refers to the family of AI techniques (using artificial neural networks with many layers) that powers today's image recognition, speech assistants, language models like ChatGPT, and more. The list was compiled as a "top 100" starting point for anyone entering the AI research field or wanting to understand the foundational ideas behind modern AI. Rather than overwhelming you with thousands of papers, the author hand-picked the most cited and impactful ones, organized by topic area: things like how neural networks are trained, how they understand images, how they process language, and how they can be applied to robotics and games. This is primarily useful to researchers, students, or technically curious founders who want to understand the intellectual history of AI, why these systems work the way they do, and what the key breakthroughs were. It's not a tool you run; it's a reference document with links to downloadable research papers. Important note: the list stopped being updated after 2017 because the volume of new AI papers became impossible to curate manually. So while it's an excellent historical foundation covering the era that established deep learning, it doesn't include the transformer architecture papers (like the one behind ChatGPT) or anything from the explosion of generative AI research that followed.