# VAA-CSEC: Vote-guided Advantage Allocation for Chinese Semantic Error Correction

arXiv: https://arxiv.org/abs/2609.36804 · HTML: https://mlsanity.com/paper/2609.36804/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Reinforcement learning for reasoning, one of 17 directions in LLM reasoning.
- Builds on 15 of its 36 references that are in the map.
- First posted on arXiv 29 Sep 2026.

## Placed in

- LLM reasoning / [Reinforcement learning for reasoning](https://mlsanity.com/lr/reinforcement-learning-for-reasoning/)

## Builds on (most cited first, 15 of 15)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [mT5: A massively multilingual pre-trained text-to-text transformer](https://mlsanity.com/paper/2010.11934/) (2020)
- [GECToR -- Grammatical Error Correction: Tag, Not Rewrite](https://mlsanity.com/paper/2005.12592/) (2020)
- [The Unreliability of Explanations in Few-shot Prompting for Textual Reasoning](https://mlsanity.com/paper/2205.03401/) (2022)
- [A Simple Recipe for Multilingual Grammatical Error Correction](https://mlsanity.com/paper/2106.03830/) (2021)
- [EDGE-GRPO: Entropy-Driven GRPO with Guided Error Correction for Advantage Diversity](https://mlsanity.com/paper/2507.21848/) (2025)
- [SynGEC: Syntax-Enhanced Grammatical Error Correction with a Tailored GEC-Oriented Parser](https://mlsanity.com/paper/2210.12484/) (2022)
- [CEC-Zero: Zero-Supervision Character Error Correction with Self-Generated Rewards](https://mlsanity.com/paper/2512.23971/) (2025)
- [When Right Meets Wrong: Bilateral Context Conditioning with Reward-Confidence Correction for GRPO](https://mlsanity.com/paper/2603.13134/) (2026)
- [NaSGEC: a Multi-Domain Chinese Grammatical Error Correction Dataset from Native Speaker Texts](https://mlsanity.com/paper/2305.16023/) (2023)
- [LLMCL-GEC: Advancing Grammatical Error Correction with LLM-Driven Curriculum Learning](https://mlsanity.com/paper/2412.12541/) (2024)
- [DPCSpell: A Transformer-based Detector-Purificator-Corrector Framework for Spelling Error Correction of Bangla and Resource Scarce Indic Languages](https://mlsanity.com/paper/2211.03730/) (2022)
- [CSED: A Chinese Semantic Error Diagnosis Corpus](https://mlsanity.com/paper/2305.05183/) (2023)

## Abstract (arXiv)

Chinese Semantic Error Correction (CSEC) targets semantic errors in Chinese text, which are typically more subtle and complex than spelling and grammatical errors but remain relatively underexplored. Existing LLM-based approaches face two recurring obstacles in this task: over-correction, and unclear interaction between Chain-of-Thought (CoT) reasoning and self-consistency decoding, such that the benefits brought by CoT cannot be reliably transferred to final corrections. We propose Vote-guided Advantage Allocation for CSEC (VAA-CSEC), a multi-stage framework that combines CoT distillation, Supervised Fine-Tuning (SFT), Reinforcement Learning (RL) and self-consistency decoding. During RL, we design a task-specific reward function that directly aligned with the minimal-editing principle of CSEC. We further introduce Group-Level Relative Policy Optimization (GLPO), which reallocates GRPO advantages according to the margin between individual rollout rewards and the vote-aggregated group reward, aligning the RL training objective with the self-consistency objective used at inference time. Experiments on CSED-C and NaSGEC-Exam show that VAA-CSEC outperforms all LLM-based baselines on CSED-C with an F0.5 of 47.72%, achieves the highest recall of 42.15% among all methods, and establishes a new state of the art of 41.55% F0.5 on NaSGEC-Exam.