Agriculture voice assistant speech is better with modular processing pipeline
Model-Agnostic and Language-Agnostic Voice Pipeline Improvement for the Agriculture Domain
Artificial IntelligenceComputation and LanguageSound
Summary
Farmers often use voice to get advice about farming, but background noise and special farming words confuse speech recognition software. The authors created a step-by-step process that cleans the audio, picks out the farmer's voice, and fixes words using farming-specific lists. This process works with many speech recognition systems without changing them and improves accuracy, especially when multiple people are talking. Their method helps machines understand farmers' questions more clearly in several Indian languages.
What this means in practice
- •For voice assistant developers: Improve speech recognition accuracy for agricultural queries by adding modular preprocessing and domain-aware corrections without changing ASR models.
- •For call center technology teams: Enhance transcription quality for conversations with background noise and multiple speakers by using target-speaker selection and domain-specific lexicons.
Authors
Aakash Singh, Lakshmi Pedapudi, Chandrashekar M S, Sanyam Singh, Naga Ganesh, Vineet Singh
Abstract
FarmerChat is Digital Green's AI-powered agricultural advisory assistant for smallholder farmers, who access it in their own language through text, voice, or photographs. Voice is a critical channel for this population, yet field-recorded speech is challenging for general-purpose automatic speech recognition (ASR) because recordings frequently contain machinery noise, background media, competing speakers, and domain-specific agricultural vocabulary. These conditions disproportionately affect crop, pest, chemical, and quantity terms that carry the meaning of a farmer's query. We present a modular, model-agnostic pipeline for improving ASR quality in FarmerChat without fine-tuning or replacing the underlying ASR model. The pipeline combines gated audio enhancement, speaker diarization and target-speaker selection, ASR, domain-aware correction using a weighted agricultural lexicon, and a quality gate for detecting unreliable transcripts. Only the diarization stage is fine-tuned; all other stages use off-the-shelf models behind common interfaces. We evaluate the pipeline on human-annotated FarmerChat recordings in Hindi, Telugu, and Odia using word error rate (WER) and a domain-weighted error rate that gives greater importance to agricultural terminology. The largest improvements occur on multi-speaker recordings, where target-speaker selection prevents competing speech from entering the transcript. Across the full corpus, the pipeline reduces WER by 16-23% relative on three cloud ASR models and by 5% on an on-device model. On multi-speaker recordings, the reductions are 32-42% for the cloud models and 16% for the on-device model. All reported reductions are statistically significant. These results show that targeted preprocessing, speaker selection, and domain-aware post-processing can substantially improve agricultural speech transcription while preserving the underlying ASR model.