ACL Findings2026
Rolling Out Data Quality Overnight, without losing the plot: A Multi-Agent System for Speech Data Quality Management
Rishabh Kumar, Abhinav Painuli, Chriss Philip Saji, Devesh Soni, Amrith Krishna, Ganesh Ramakrishnan
Abstract
Quality management when creating large-scale speech datasets is essential for building reliable downstream models, yet verification pipelines are often brittle, domain-specific, and expertiseintensive. We introduce SpeechQM-Agent, a natural language-driven agentic framework that compiles user requirements into dependencyaware DAG workflows over modular tools for audio, transcript, and metadata verification. A central planner LLM enforces prerequisites and supports execution-time replanning (e.g., rerunning failed steps or swapping tools), reducing manual pipeline engineering and improving robustness across heterogeneous vendor formats and multilingual settings. We also release SpeechQM-Dataset, a multilingual benchmark with controlled, vendor-inspired quality artifacts spanning 24 verification tasks. Across experiments, SpeechQM-Agent attains 80-90% agreement with expert verification while requiring <20% of the cost and time of manual QC, and we further validate transfer to real vendorsupplied corpora. Planner LLM comparisons highlight fidelity-efficiency trade-offs.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
Builds on12
- Plug-and-Play Policy Planner for Large Language Model Powered Dialogue AgentsYang Deng, Wenxuan Zhang, Wai Lam, See-Kiong Ng et al.ICLR 2024 · 86 citations
- HAIChart: Human and AI Paired Visualization SystemYupeng Xie, Yuyu Luo, Guoliang Li, Nan TangVLDB 2024 · 47 citations
- Automatic Speech Recognition in Sanskrit: A New Speech Corpus and Modelling InsightsDevaraja Adiga, Rishabh Kumar, Amrith Krishna, Preethi Jyothi et al.ACL Findings 2021 · 24 citations
- DFA-RAG: Conversational Semantic Router for Large Language Model with Definite Finite AutomatonYiyou Sun, Junjie Hu, Wei Cheng, Haifeng ChenICML 2024 · 4 citations
- AFlow: Automating Agentic Workflow GenerationJiayi Zhang, Jinyu Xiang, Zhaoyang Yu, Fengwei Teng et al.ICLR 2025
Related papers
- SpeechLLM-as-Judges: Towards General and Interpretable Speech Quality EvaluationHui Wang, Jinghua Zhao, Yifan Yang, Shujie Liu et al.ACL 2026 · 21 citations
- PAT-Agent: Autoformalization for Model CheckingXinyue Zuo, Yifan Zhang, Hongshu Wang, Yufan Cai et al.ASE 2025 · 1 citation
- Aligning Audio-Visual Joint Representations with an Agentic WorkflowShentong Mo, Yibing SongNeurIPS 2024 · 7 citations
- MEnvAgent: Scalable Polyglot Environment Construction for Verifiable Software EngineeringChuanzhe Guo, Jingjing Wu, Sijun He, Yang Chen et al.ICML 2026 · 3 citations
- BFCL Audio: An Audio Function Calling Evaluation for Large Language ModelsHuanzhi Mao, Aditya Ghai, Imra Dawoodani, Tony Ginart et al.ICML 2026
