Improving Neural Text Style Transfer by Introducing Loss Function Sequentiality.
Chinmay RaneGaël DiasAlexis LechervyAsif EkbalPublished in: SIGIR (2021)
Keyphrases
- loss function
- logistic regression
- pairwise
- empirical risk
- learning to rank
- support vector
- risk minimization
- information retrieval
- boosting framework
- regularization term
- reproducing kernel hilbert space
- hinge loss
- text mining
- stochastic gradient descent
- convex loss functions
- text documents
- markov random field
- transfer learning
- special case
- solution path