Learning Neural Sequence-to-Sequence Models from Weak Feedback with Bipolar Ramp Loss.
Laura JehlCarolin LawrenceStefan RiezlerPublished in: CoRR (2019)
Keyphrases
- prior knowledge
- hidden state
- learning rules
- learning process
- markov models
- probabilistic model
- learning models
- learned models
- learning algorithm
- learning tasks
- accurate models
- unsupervised learning
- online learning
- action sequences
- supervised learning
- biologically plausible
- structured prediction
- hidden markov models
- active learning
- neural network
- spiking neural networks
- learning problems
- statistical models
- loss function
- learning systems
- markov random field
- relevance feedback
- semi supervised
- reinforcement learning