Provably Neural Active Learning Succeeds via Prioritizing Perplexing Samples.
Dake BuWei HuangTaiji SuzukiJi ChengQingfu ZhangZhiqiang XuHau-San WongPublished in: CoRR (2024)
Keyphrases
- active learning
- sample selection
- training set
- representative samples
- training examples
- network architecture
- neural network
- stratified sampling
- machine learning
- data sets
- training samples
- learning algorithm
- active sampling
- selective sampling
- neural model
- bio inspired
- random sampling
- learning strategies
- supervised learning
- worst case
- semi supervised
- associative memory
- relevance feedback
- experimental design
- decision trees
- neural fuzzy
- active learning strategies
- multi class
- music emotion classification
- pool based active learning
- training data
- learning environment
- feature space
- batch mode
- software engineering
- cost sensitive