A field guide to the GAN research explosion

What it does This repository is a manually curated bibliography of generative adversarial network papers, organized into application areas like image translation, facial attribute manipulation, and super-resolution, plus theory, machine learning techniques, and interdisciplinary topics such as medicine and music. Each entry typically pairs the paper with an arXiv or project link and, when available, a link to the corresponding implementation. It is essentially a reading list and lookup table rather than a software project.

The interesting bit The list attempts to be exhaustive rather than selective, covering everything from the original 2014 Goodfellow paper to CVPR 2020 oral presentations, which makes it a decent proxy for watching how a single architecture family swallowed entire research subfields. The maintainer also includes their own TensorFlow reimplementations for some entries, blurring the line between bibliography and personal lab notebook.

Key highlights

• Covers more than a dozen application domains, including niche areas like gaze correction, shadow removal, and makeup transfer.
• Links papers to code repositories where possible, saving the usual hunt through arXiv footnotes.
• Tracks theoretical advances and training improvements (Wasserstein GANs, conditional adversarial training) alongside applications.
• Includes tutorials, blog posts, and interdisciplinary work, not just pure computer vision.
• Maintained since at least 2016, with entries running through 2020.

Caveats

• The README is utilitarian and occasionally rough: the opening description misspells “Generative” as “Generation,” and the markdown formatting is inconsistent.
• Some category links in the table of contents appear broken or point to incorrect anchors (e.g., multiple sections link to #Facial-Attribute-Manipulation).
• It is a static list, not a searchable database, so finding a specific paper means scrolling through long, single-page sections.

Verdict Worth bookmarking if you are doing a literature review on GANs or trying to find reference implementations for a specific image-to-image task. Skip it if you are looking for a framework, a training script, or critical commentary on paper quality.