A curated list of resources on continual zero-shot learning, inspired by awesome-computer-vision.
Please feel free to send me pull requests or email (williamyi96@gmail.com) to add links.
- DVGR: Dynamic VAEs with Generative Replay for Continual Zero-shot Learning. Subhankar Ghosh. [paper] [code]
- MCZSL: Vinay Kumar Verma, Kevin Liang, Nikhil Mehta, Lawrence Carin. Meta-Learned Attribute Self-Gating for Continual Generalized Zero-Shot Learning. arXiv (2021). [arXiv]
- Generalized Continual Zero-Shot Learning / Generative Replay-based Continual Zero-Shot Learning. Chandan Gautam, Sethupathy Parameswaran, Ashish Mishra, Suresh Sundaram. [paper1] [paper2]
- Ghost: Arthur Douillard, Eduardo Valle, Charles Ollion, Thomas Robert, Matthieu Cord. "Insights from the Future for Continual Learning", CVPR Continual Learning Workshop 2021, [arxiv] [code]
- CN-ZSL: Ivan Skorokhodov, Mohamed Elhoseiny. "Class Normalization for Zero-Shot Learning". ICLR (2021). [page] [code]
- A-GEM Arslan Chaudhry, Marc’Aurelio Ranzato, Marcus Rohrbach, Mohamed Elhoseiny. "Efficient Lifelong Learning with A-GEM". ICLR (2019). [paper] [code]
- CN-ZSL CUB: 200 classes randomly split into 10 tasks with 20 classes per task. [data]
- CN-ZSL SUN: 717 classes randomly split into 15 tasks, the first 3 tasks have 47 classes and the rest of them have 48 classes each. [data]
- A-GEM Split CUB: 200 classes split into 20 disjoint subsets of classes.
- A-GEM Split AWA: 50 animal categories, where each task is constructed by sampling 5 classes with replacement from the total 50 classes, constructing 20 tasks.
- LZSL: Sequential training on aPY, AWA1, CUB and SUN seen classes, then evaluate on unseen classes.
- Arthur Douillard's IncLearn to do Continual Zeroshot on AwA2, APY, CUB200, and LAD.
License
To the extent possible under law, Kai Yi has waived all copyright and related or neighboring rights to this work.