Title | ||
---|---|---|
Automatic Gain Control And Multi-Style Training For Robust Small-Footprint Keyword Spotting With Deep Neural Networks |
Abstract | ||
---|---|---|
We explore techniques to improve the robustness of small-footprint keyword spotting models based on deep neural networks (DNNs) in the presence of background noise and in far-field conditions. We find that system performance can be improved significantly, with relative improvements up to 75% in far-field conditions, by employing a combination of multi-style training and a proposed novel formulation of automatic gain control (AGC) that estimates the levels of both speech and background noise. Further, we find that these techniques allow us to achieve competitive performance, even when applied to DNNs with an order of magnitude fewer parameters than our baseline. |
Year | Venue | Keywords |
---|---|---|
2015 | 2015 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING (ICASSP) | keyword spotting, automatic gain control, multi-style training, small-footprint models |
Field | DocType | ISSN |
Background noise,Pattern recognition,Noise measurement,Computer science,Speech recognition,Robustness (computer science),Keyword spotting,Footprint,Artificial intelligence,Automatic gain control,Deep neural networks | Conference | 1520-6149 |
Citations | PageRank | References |
7 | 0.58 | 5 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Rohit Prabhavalkar | 1 | 163 | 22.56 |
Raziel Álvarez | 2 | 30 | 3.84 |
Carolina Parada | 3 | 242 | 13.11 |
Preetum Nakkiran | 4 | 64 | 6.05 |
Tara N. Sainath | 5 | 3497 | 232.43 |