ConBERT-RL: A policy-driven deep reinforcement learning based approach for detecting homophobia and transphobia in low-resource languages.
Vivek Suresh RajChinnaudayar Navaneethakrishnan SubalalithaLavanya Sambath KumarFrank G. GlavinBharathi Raja ChakravarthiPublished in: Nat. Lang. Process. J. (2024)
Keyphrases
- reinforcement learning
- optimal policy
- rl algorithms
- policy search
- action selection
- markov decision process
- state space
- actor critic
- action space
- reinforcement learning algorithms
- reinforcement learning problems
- function approximation
- state and action spaces
- markov decision processes
- control policy
- allocation policies
- model free
- partially observable
- markov decision problems
- control policies
- policy evaluation
- total reward
- expressive power
- function approximators
- partially observable domains
- partially observable markov decision processes
- average reward
- state action
- policy gradient
- learning algorithm
- reward function
- reinforcement learning methods
- continuous state
- policy iteration
- learning problems
- dynamic programming
- agent learns
- partially observable environments
- agent receives
- model free reinforcement learning
- finite state
- temporal difference methods
- resource allocation
- decision problems
- semi markov decision process
- approximate dynamic programming
- optimal control
- temporal difference
- control problems
- approximate policy iteration
- multi agent
- infinite horizon
- temporal difference learning
- exploration exploitation tradeoff
- transition model
- machine learning
- learning process
- fully observable
- partially observable markov decision process
- long run
- single agent