data-centric-AI

Why we included this project

Researchers and ML engineers who care about where models actually come from will find this a useful starting point. Rather than a code library you install, it is a curated index of papers, tutorials, and surveys, with pointers to related tools and code, built around the idea that improving training data quality and structure can matter as much as changing the model architecture. The maintainer has deliberately picked works that represent distinct ideas instead of trying to be exhaustive, which makes the list easier to navigate than a raw paper dump. It also threads in related resources like graph structure learning benchmarks and discussions of data-centric concepts behind models such as GPT and Segment Anything, helping newcomers see how these ideas connect. Each entry links out to its source, so it works well as a literature map when you are scoping a project or building a reading list.