KIRA：面向儿童的双通路分析全栈阅读筛查系统

arXiv cs.CL 论文
speech-recognition reading-screening education child-speech phoneme-detection full-stack-system

摘要

KIRA 是一个面向儿童早期阅读筛查的全栈系统，结合词级流利度评分与音素级错误检测的双通路分析，将儿童语音识别的词错误率从 13.88% 降至 8.84%、音素错误率从 41.66% 降至 23.00%，并配套教师端 Web 界面以支持课堂部署与针对性干预。

arXiv:2610.28840v1 Announce Type: new Abstract: Early-grade reading screening is critical for identifying children at risk of reading difficulty. Existing AI-based screeners are often limited in the scope of their analyses and evaluation metrics, or are proprietary systems that lack transparent model specifications and evaluation. We present KIRA, a dual-pathway screening system that pairs word-level fluency scoring with detailed phoneme-level error detection. We adapt pretrained word and phoneme recognizers to children's speech, reducing word error rate from 13.88% to 8.84% and phoneme error rate from 41.66% to 23.00%. Powered by these child-adapted models, we implement Word Reading (WRE) and Oral Reading Fluency (ORF), two foundational tasks in K-2 screening. The system integrates these recognizers with a web-based teacher-facing interface presenting fluency scores alongside phoneme feedback for potential targeted intervention. KIRA brings modern child-speech recognition into a full-stack, deployable, classroom-ready reading screener.
查看原文
查看缓存全文

缓存时间: 2026/10/22 16:21 

# KIRA: A Full-Stack Reading Screening System with Dual-Pathway Analysis for Children
Source: [https://arxiv.org/abs/2610.28840](https://arxiv.org/abs/2610.28840)
[View PDF](https://arxiv.org/pdf/2610.28840)

> Abstract:Early\-grade reading screening is critical for identifying children at risk of reading difficulty\. Existing AI\-based screeners are often limited in the scope of their analyses and evaluation metrics, or are proprietary systems that lack transparent model specifications and evaluation\. We present KIRA, a dual\-pathway screening system that pairs word\-level fluency scoring with detailed phoneme\-level error detection\. We adapt pretrained word and phoneme recognizers to children's speech, reducing word error rate from 13\.88% to 8\.84% and phoneme error rate from 41\.66% to 23\.00%\. Powered by these child\-adapted models, we implement Word Reading \(WRE\) and Oral Reading Fluency \(ORF\), two foundational tasks in K\-2 screening\. The system integrates these recognizers with a web\-based teacher\-facing interface presenting fluency scores alongside phoneme feedback for potential targeted intervention\. KIRA brings modern child\-speech recognition into a full\-stack, deployable, classroom\-ready reading screener\.

## Submission history

From: Maggie Dong \[[view email](https://arxiv.org/show-email/a089a4e2/2610.28840)\] **\[v1\]**Tue, 20 Oct 2026 19:10:45 UTC \(579 KB\)

相似文章

Show HN：开源的早期阅读器，提供自然拼读辅助

Hacker News Top

Show HN：Yara Reads 是一个开源的儿童早期阅读应用，专为刚学完自然拼读 phonics 的孩子设计，每个单词都可逐音素拆解朗读，并内置按发音校验词汇的检查器。项目基于 Cloudflare Worker 构建，支持通过 MCP 接入 Claude/ChatGPT 等代理自动生成符合孩子音素水平的故事。

面向AI教育中算法追踪与问题求解的检索增强型辅导

arXiv cs.AI

本文介绍了KITE，一个基于检索增强生成（RAG）的智能辅导系统，用于AI教育中的算法推理和问题求解。该系统采用意图感知的苏格拉底式回应策略和多模态RAG，提供基于课程内容、符合教学法的反馈，并通过指标评估、专家评审和模拟学生交互进行评价。

车载对话助手多轮对话的自动化评测

arXiv cs.CL

研究人员提出一个用于评测车载对话助手（ICA）多轮对话能力的自动化黑盒测试框架，采用策略引导的用户模拟器、对抗性策略管理器和两级 LLM 评审器，在工业级 ICA 上验证该方法与人类标注高度一致，并发现 2.96 倍的独特失败类型。

MIRA：用于对齐文本到音乐生成与用户意图的音乐意图细化智能体

Hugging Face Daily Papers

本文提出 MIRA（Musical Intent Refinement Agent），一种在测试时通过将用户意图细化为可验证 rubric 并进行轨迹感知树搜索来改进文本到音乐生成对齐的智能体，并发布了由音乐专家标注的基准 MuRA-Bench，使开源生成器达到接近 Suno、Mureka 等商业系统的性能。