Discrete Prompt Compression with Reinforcement Learning.
Hoyoun JungKyung-Joong KimPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- function approximation
- data compression
- continuous state
- continuous domains
- compression scheme
- compression algorithm
- model free
- continuous state and action spaces
- compression ratio
- discrete version
- markov decision processes
- state space
- transfer learning
- optimal policy
- image compression
- robotic control
- temporal difference
- reinforcement learning algorithms
- continuous state spaces
- discrete geometry
- multi agent reinforcement learning
- supervised learning
- lossless compression
- compression rate
- sequence prediction
- data sets
- machine learning