Neural Feature-Adaptation for Symbolic Predictions Using Pre-Training and Semantic Loss.
Vedant ShahAditya AgrawalLovekesh VigAshwin SrinivasanGautam ShroffTanmay T. VerlekarPublished in: CoRR (2022)
Keyphrases
- connectionist models
- neural learning
- image features
- high level
- neural network
- semantic information
- covariate shift
- semantic knowledge
- natural language
- recurrent networks
- semantic representation
- network architecture
- training algorithm
- domain specific
- training process
- knowledge base
- multiple features
- semantic network
- semantic similarity
- test set
- training examples
- semantic web
- online learning
- domain ontology
- low level
- training set
- semantic description
- training samples
- decision trees