A Bi-objective Perspective on Controllable Language Models: Reward Dropout Improves Off-policy Control Performance.
Changhun LeeChiehyeon LimPublished in: CoRR (2023)
Keyphrases
- language model
- bi objective
- language modeling
- multi objective
- efficient solutions
- n gram
- document retrieval
- statistical language models
- probabilistic model
- speech recognition
- language modelling
- multiple objectives
- query terms
- retrieval model
- query expansion
- multi objective optimization
- test collection
- vector space model
- ad hoc information retrieval
- information retrieval
- knapsack problem
- smoothing methods
- document ranking
- language model for information retrieval
- context sensitive
- reinforcement learning
- relevance model
- pseudo relevance feedback
- evolutionary algorithm
- statistical language modeling
- shortest path problem
- network design
- okapi bm
- language models for information retrieval
- bayesian networks