Annotating Panic in Social Media using Active Learning, Transformers and Domain Knowledge.
Sandra MitrovicFabio FrisoneSuryam GuptaChiara LuciforaDragana CarapicCarlo SchillaciSamuele Di GiovanniAyushi SinghPublished in: ICDM (Workshops) (2023)
Keyphrases
- domain knowledge
- active learning
- social media
- social networking
- semi supervised
- metadata
- domain experts
- selective sampling
- social networks
- random sampling
- knowledge sources
- training examples
- learning strategies
- background knowledge
- labeled data
- machine learning
- transfer learning
- social media data
- learning algorithm
- prior knowledge
- learning process
- supervised learning
- training set
- unlabeled data
- explicit knowledge
- domain theory
- experimental design
- active learning strategies
- big data
- semantic annotation
- domain ontology
- pool based active learning
- online social networks
- knowledge base
- cost sensitive
- user generated content
- image retrieval
- domain specific knowledge
- manual annotation
- crisis management
- relevance feedback
- prior domain knowledge
- neural network
- music emotion classification
- batch mode
- information diffusion
- e learning
- reinforcement learning
- semantic information