Picovoice/

speech-to-intent-benchmark

Made in Vancouver, Canada by Picovoice

This framework benchmarks the accuracy of Picovoice's Speech-to-Intent engine, Rhino. It compares the accuracy of Rhino with: Amazon Lex Google Dialogflow IBM Watson Microsoft LUIS

Command acceptance rate is the probability of an engine correctly understanding the spoken command. Below is the summary:

The figure below depicts engines performance at each SNR:

The speech data… See the full description on the dataset page: https://huggingface.co/datasets/Picovoice/speech-to-intent-benchmark.