AmpliBias: Mitigating Dataset Bias through Bias Amplification in Few-shot Learning for Generative Models.
Donggeun KoDongjun LeeNamjun ParkKyoungrae NohHyeonjin ParkJaekwang KimPublished in: CIKM (2023)
Keyphrases
- generative model
- discriminative learning
- learning algorithm
- prior knowledge
- mixture model
- boltzmann machine
- hidden variables
- learning process
- machine learning
- semi supervised
- language model
- deep belief networks
- probabilistic model
- reinforcement learning
- parameter learning
- object categories
- representational power
- deep learning
- computer vision
- object categorization
- learning tasks
- unsupervised learning
- active learning
- information extraction