发表机构
CAKE by VPBank（VPBank旗下CAKE）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出SEAR方法，通过片段证据感知的数据生成与路由，将弱文本可答和强音频依赖的MCQ分别用于监督微调和强化学习，在MLC-SLM挑战赛任务2中达到90.92%准确率。
AI 中文摘要
本文描述了我们在第二届多语言对话语音语言模型（MLC-SLM）挑战赛任务2中的系统。我们采用片段证据感知的数据和后训练流程来适配Qwen3-Omni-30B-A3B-Instruct模型。一个语言模型将带时间戳的自动语音识别（ASR）结果转换为连贯的事件片段，这些片段通过边界余量进行扩展，并从原始录音中裁剪出来。随后，我们使用Qwen3.6-27B合成互补的语义多项选择题（MCQs），并使用Gemini 3.1 Flash-Lite合成声学多项选择题，接着进行结构、接地、答案一致性和目标模型可训练性检查，最终在21种语言和口音变体中生成359,825个经过验证的多项选择题。一个仅文本的探针将数据划分为弱项（文本可回答）用于监督微调，以及强项（依赖音频）用于通过组序列策略优化（GSPO）进行强化学习，并通过去偏优势、序列级重要性校正和动态过滤来稳定训练。我们的系统在最终官方评估集上达到了90.92%的准确率。
英文摘要
This paper describes our system for Task~2 of the second Multilingual Conversational Speech Language Model (MLC-SLM) Challenge. We adapt Qwen3-Omni-30B-A3B-Instruct with a segment-evidence-aware data and post-training pipeline. A language model converts timestamped ASR into coherent event spans, which are expanded by a boundary margin and cropped from the original recording. We then synthesize complementary semantic MCQs with Qwen3.6-27B and acoustic MCQs with Gemini~3.1 Flash-Lite, followed by structural, grounding, answer-consistency, and target-model trainability checks, yielding 359,825 verified MCQs across 21 language and accent variants. A text-only probe partitions the data into weak, text-answerable items used for supervised fine-tuning and strong, audio-dependent items used for reinforcement learning with Group Sequence Policy Optimization (GSPO), stabilized by debiased advantages, sequence-level importance correction, and dynamic filtering. Our system obtains 90.92% accuracy on the final official evaluation set.
