FANDOM


  • Reinforcement learning
  • Type-supervised learning
  • Semisupervised learning
  • Multiple-instance learning: "uses training examples grouped into sets, commonly referred to as bags. A single label (either positive or negative) is assigned to each bag. A positive label indicates that the bag includes at least one example belonging to the positive class. A negative label denotes that all instances of the bag belong to the negative class." (Torresani, 2016)

TODO: ACL Tutorial 2013 on semantic parsing

What kind of supervision is available?

  • Annotated parse trees [Miller et al. 1994]
  • Sentence-LF pairs [Zettlemoyer and Collins 2005]
  • Question-answer pairs [Clarke et al. 2010]
  • Instruction-demonstration pairs [Chen and Mooney 2011]
  • Conversation logs [Artzi and Zettlemoyer 2011]
  • Visual sensors [Matuszek et al. 2012a]