用中文家长会录音自动生成个性化教育计划，兼顾隐私与效率
Automated IEP Generation from Traditional Chinese Parent-Teacher Interviews via Corpus-Grounded Feature Diffusion
- 基于语料特征扩散，从少量专家对话中提炼生成模板
- 本地化推理下生成质量超多个主流模型，且速度更快
- 适合特殊教育、教育科技从业者及关注数据安全的团队
撰写个性化教育计划（IEP）是一项高人力、高知识密度的工作；英文研究显示生成式AI可显著缩短起草时间，但繁体中文环境下的自动化IEP生成仍几乎空白，受限于领域数据稀缺、严格隐私法规及缺乏本地评估基准。本文提出一种低资源微调流程——语料引导特征扩散（CGFD）：(1) 通过τ阈值与标志感知得分上限，从25份双专家高分转录稿中筛选种子样本；(2) 提取特征轮廓（句长、结构、量化模板），结合口语化采样多样性控制注入LLM提示，驱动扩散生成；(3) 使用15份专家黄金样本作为扩散锚点，目标生成585份样本，最终获得567个有效样本，构建582样本训练集，用于在Breeze-7B上采用QLoRA微调；(4) 推理时采用语法约束解码（GCD）强制执行层级SMART目标阶梯结构。在55样本的结构压力测试中发现意外结果：在繁体中文词元预算下，GCD反而降低性能——无GCD路径以34%更低的中位延迟实现100%结构通过率，优于含GCD方案。在n=10正式留出测试集上，无GCD推理路径达到BERTScore F1 = 0.779，超越GPT-5.4（0.726）、DeepSeek-V3.2（0.703）、Gemini-3-Flash-Preview（0.703）和Llama-4-Maverick（0.700）的零样本基线，同时保持完全本地、隔离网络推理。该系统填补了繁体中文特殊教育自然语言处理的空白，提供一种可扩展、隐私保护的本地推理解决方案。
原文摘要 · Abstract (English)
Writing Individualized Education Programs (IEPs) is a high-labor, knowledge-intensive document burden; English-language research has demonstrated that generative AI can significantly reduce drafting time, yet automated IEP generation in Traditional Chinese remains virtually unexplored due to domain data scarcity, strict privacy regulations, and the absence of local evaluation benchmarks. We propose a low-resource fine-tuning pipeline centered on Corpus-Grounded Feature Diffusion (CGFD): (1) 25 dual-expert high-score seed transcripts are selected via a tau threshold with flag-aware score caps; (2) a FeatureProfile (sentence length, structure, quantification templates) is extracted from seeds and injected into LLM prompts alongside Verbalized-Sampling-style diversity control to drive diffusion; (3) 15 expert gold seeds are used as diffusion anchors, targeting 585 samples; 567 valid diffusion samples are obtained, yielding a 582-sample training set used to fine-tune Breeze-7B with QLoRA; (4) schema-constrained inference via Grammar-Constrained Decoding (GCD) enforces a hierarchical SMART Goal Ladder schema at inference time. Ablation results on a 55-sample schema stress set reveal an unexpected finding: GCD is counterproductive under Traditional Chinese token budgets -- the no-GCD path achieves 100% schema pass rate at 34% lower median latency, outperforming GCD on both reliability and speed. On the n=10 formal hold-out, the no-GCD inference path achieves BERTScore F1 = 0.779, exceeding GPT-5.4 (0.726), DeepSeek-V3.2 (0.703), Gemini-3-Flash-Preview (0.703), and Llama-4-Maverick (0.700) zero-shot baselines while maintaining fully local, air-gapped inference. This system addresses a gap in Traditional Chinese special-education NLP and offers a scalable, privacy-preserving local inference solution under an industrial engineering paradigm.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。