Progressive Voice Trigger Detection: Accuracy vs Latency
| Autor: | Erik Marchi, Hywel Richards, John Bridle, Vineet Garg, Pascal Clark, Siddharth Sigtia |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2020 |
| Předmět: |
FOS: Computer and information sciences
Sound (cs.SD) Computer Science - Machine Learning Phrase Exploit Computer science Speech recognition Computer Science - Human-Computer Interaction Context (language use) Computer Science - Sound Human-Computer Interaction (cs.HC) Machine Learning (cs.LG) Audio and Speech Processing (eess.AS) False rejection rate Test set FOS: Electrical engineering electronic engineering information engineering Latency (engineering) Electrical Engineering and Systems Science - Audio and Speech Processing |
| Zdroj: | ICASSP |
| Popis: | We present an architecture for voice trigger detection for virtual assistants. The main idea in this work is to exploit information in words that immediately follow the trigger phrase. We first demonstrate that by including more audio context after a detected trigger phrase, we can indeed get a more accurate decision. However, waiting to listen to more audio each time incurs a latency increase. Progressive Voice Trigger Detection allows us to trade-off latency and accuracy by accepting clear trigger candidates quickly, but waiting for more context to decide whether to accept more marginal examples. Using a two-stage architecture, we show that by delaying the decision for just 3% of detected true triggers in the test set, we are able to obtain a relative improvement of 66% in false rejection rate, while incurring only a negligible increase in latency. Camera Ready Version: ICASSP 2021 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|