RLDE: A Reinforcement Learning Framework for Interpretable Iterative Document Editing Researchers introduced RLDE, a reinforcement learning framework that treats long-document revision as a reference-free Markov decision process using a limited action space of NOOP, INSERT, DELETE and REPLACE, with Pointer Networks selecting token-level spans and a frozen BART-base model generating suggestions. On the BEA-2019 dataset, RLDE produced statistically significant gains (p < 0.01) across all metrics, raising Flesch Reading Ease from 49.8 to 59.4 and lowering perplexity from 63.7 to 39.8 while preserving 93% of semantic information, with human raters scoring edits 4.9/5 for meaning preservation and 4.4/5 for edit quality. The authors report RLDE outperforms Chain-of-Thought prompt-engineered models including GPT-4 and LLaMA-2 at very low computational inference cost. arXiv:2611.22039v1 Announce Type: new Abstract: Improving the quality, clarity, and flow of long documents remains a challenging problem in NLP, as human editing typically occurs through iterative refinement rather than a single-pass rewrite. Most existing automated systems act as black-box generators or rely on supervised learning, limiting interpretability and failing to capture the sequential nature of revisions. In this paper, we propose RLDE, an automated end-to-end metric-driven RL framework that formulates document revision as a reference-free MDP problem unsupervised with respect to paired document-revision targets . Working with a limited action space NOOP, INSERT, DELETE, REPLACE , RLDE leverages Pointer Networks for dynamic span selection at the token level and sends generative actions to the frozen BART-base model to obtain context-aware suggestions. PPO agents learn to optimize a multi-objective reward function based on readability, perplexity, and grammar metrics with the help of a novel SBERT-based Semantic Preservation Penalty constraint to avoid deletion degeneracy. Evaluations on the BEA-2019 dataset showed statistically significant results p < 0.01 across all metrics, where Flesch Reading Ease increased from 49.8 to 59.4, perplexity decreased from 63.7 to 39.8, while preserving 93% of semantic information. When compared to Chain-of-Thought CoT prompt-engineered models like GPT-4 and LLaMA-2, RLDE shows high performance at a very low computational inference cost. Moreover, the quality of edits generated by RLDE has been verified via human evaluations, scoring 4.9/5 for meaning preservation and 4.4/5 for edit quality.