InfoPrompt: Information-Theoretic Soft Prompt Tuning for Natural Language Understanding.
Junda WuTong YuRui WangZhao SongRuiyi ZhangHandong ZhaoChaochao LuShuai LiRicardo HenaoPublished in: CoRR (2023)
Keyphrases
- information theoretic
- natural language understanding
- information theory
- mutual information
- text understanding
- theoretic framework
- knowledge representation
- semantic analysis
- language understanding
- natural language
- natural language processing
- semantic representations
- information bottleneck
- jensen shannon divergence
- spoken dialog systems
- information theoretic measures
- dialogue system
- log likelihood
- multi modality
- minimum description length
- relative entropy
- image segmentation
- entropy measure
- kullback leibler divergence
- artificial intelligence
- machine learning
- graphical models
- domain knowledge
- training data
- metadata
- feature selection
- computer vision