// the find
pliang279/awesome-multimodal-ml
Reading list for research topics in multimodal machine learning
A giant hand-maintained bibliography of multimodal ML papers — representation learning, fusion, pretraining, VQA, embodied navigation, and more — put together by a CMU PhD student. It's for researchers who want a map of a subfield's literature, not for anyone looking to write code this week.
The taxonomy is genuinely useful: papers are split into fusion, alignment, pretraining, co-learning, missing-modality, bias/fairness, etc., so you can find the right subfield fast instead of scrolling a flat list. A large fraction of entries link straight to the paper's code repo, so it doubles as paper-to-implementation index. It's not just link-dumping — there's a tutorial paper, CVPR/NAACL tutorial slides, and full lecture videos from an actual CMU course (11-777/11-877) behind it, so there's pedagogy attached to the list.
It's a single README with no tooling — no link checker, no script to verify the ~250 arXiv/code URLs still resolve, so bitrot is just a matter of time at this size. There's zero curation signal beyond chronological/venue order: a 2011 Boltzmann machine paper sits with equal visual weight to a 2023 ICLR paper, so a newcomer can't tell what's foundational versus niche without already knowing the field. Last push was August 2024, and the 2024 entries are thin, so it's already missing most of the recent multimodal-LLM wave (GPT-4V-era work, LLaVA-style VLMs, etc.) that a 2026 reader would actually want.