We consider the problem of performing Spoken Language Understanding (SLU) on small devices typical of IoT applications. Our contributions are twofold. First, we outline the design of an embedded, private-by-design SLU system and show that it has performance on par with cloud-based commercial solutions. Second, we release the datasets used in our experiments in the interest of reproducibility and in the hope that they can prove useful to the SLU community.

PDF Abstract


Introduced in the Paper:

Snips-SmartLights Snips-SmartSpeaker

Used in the Paper:

LibriSpeech SNIPS
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Spoken Language Understanding Snips-SmartLights Google Accuracy (%) 79.3 # 6
Spoken Language Understanding Snips-SmartLights Snips Accuracy (%) 84.2 # 5
Spoken Language Understanding Snips-SmartSpeaker Snips Accuracy-EN (%) 68.7 # 4
Accuracy-FR (%) 75.1 # 4
Spoken Language Understanding Snips-SmartSpeaker Google Accuracy-EN (%) 47.8 # 5
Accuracy-FR (%) 42.3 # 5


No methods listed for this paper. Add relevant methods here