Wireless deep learning papers, but only the ones you can run

What it does This repository is a curated bibliography mapping deep-learning papers in wireless communications to their open-source implementations. It catalogs work across physical-layer optimization, MIMO detection, channel estimation, beamforming, modulation recognition, and network resource allocation. The maintainer also runs a companion daily_arxiv project that auto-scrapes arXiv categories eess.SP and cs.IT for newly open-sourced papers.

The interesting bit The list is deliberately filtered to include only papers that actually ship code, which the maintainer notes is still uncommon in the communications field. For grad students entering a new direction without existing lab code, that scarcity makes this index less of a luxury and more of a necessity.

Key highlights

• Organized by sub-discipline, from physical-layer optimization to secure machine learning over networks
• Community-maintained: accepts additions via GitHub issues
• Companion repo uses a crawler and GitHub Actions to update daily with new open-source papers from arXiv
• Explicitly acknowledges the reproducibility gap in wireless communications research
• ~2.3k stars, suggesting it has become a de facto reference

Caveats

• The maintainer admits coverage is incomplete due to limited bandwidth
• Several roadmap items, including a list of traditional communication papers and high-citation papers without code, remain unfinished
• The primary README is in Chinese; the linked English version resides in a separate fork/repo (IIT-Lab), which may not stay in sync

Verdict Worth a bookmark for wireless comms researchers and grad students who are tired of reading about neural network detectors they can never run. Skip it if you are looking for a unified framework or toolkit.