Action Collapse Policy Gradient (ACPG) fixes the action-selection layer to a simplex ETF and claims improved discrete-action RL performance, with a theory that only covers a weighted optimal-action imitation objective.
Contrastive policy gradient: Aligning LLMs on sequence-level scores in a supervised-friendly fashion
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Imitate Optimal Policy: Prevail and Induce Action Collapse in Policy Gradient
Action Collapse Policy Gradient (ACPG) fixes the action-selection layer to a simplex ETF and claims improved discrete-action RL performance, with a theory that only covers a weighted optimal-action imitation objective.