When is Multi-task Learning Beneficial for Low-Resource Noisy Code-switched User-generated Algerian Texts?
Wafia AdouaneJean-Philippe BernardyPublished in: CodeSwitch@LREC (2020)
Keyphrases
- user generated
- multi task learning
- multi task
- multiple tasks
- gaussian processes
- social media
- learning problems
- learning tasks
- web content
- high order
- multitask learning
- user generated content
- user interests
- transfer learning
- learning algorithm
- theoretical analysis
- sentiment analysis
- social networks
- incomplete data
- inductive learning
- labeled data
- markov random field
- higher order
- knowledge representation