// the take
A README-only link list of academic survey papers covering deep learning, NLP, computer vision, recommender systems, reinforcement learning, graph learning, and embeddings. It's aimed at ML practitioners or researchers who want a starting-point survey for a subfield instead of hunting arXiv themselves.
Links are grouped by subfield and then tagged by angle (e.g. 'Transfer learning', 'Architectures') within each section, so you can scan for a specific slice of a topic rather than wading through a flat dump. Every entry includes the publication year, so you can tell at a glance whether a given 'survey' is from 2007 or 2021. Links go straight to PDFs/arXiv, no paywall or affiliate redirect layer. The sister link to eugeneyan/applied-ml for implementation write-ups shows this was built with a practitioner audience in mind, not just academics.
There's no code or tooling here — adopting it means bookmarking a README, nothing more. Last commit is March 2023, so anything on LLM alignment, RAG, diffusion models, or agent architectures — now the center of gravity in ML — is simply missing. Entries have no abstract or description, just title/year/link, so you still have to open each one to judge relevance. Several surveys are 10+ years old (2007, 2009, 2013) with no flag for which have been superseded, so the list doesn't distinguish a still-current reference from a historical one.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →