16th Annual Conference of the International Speech Communication Association

Dresden, Germany
September 6-10, 2015

Geo-Location for Voice Search Language Modeling

Ciprian Chelba, Xuedong Zhang, Keith Hall

Google, USA

We investigate the benefit of augmenting with geo-location information the language model used in speech recognition for voice-search.
    We observe reductions in perplexity of up to 15% relative on test sets obtained from both typed query data, as well as transcribed voice search data; on a subset of the test data consisting of “local” queries — search results displaying a restaurant, some address, or similar — the reduction in perplexity is even higher, up to 30% relative.
    Automatic speech recognition experiments confirm the utility of geo-location information for improved language modeling. Significant reductions in word error rate are observed both on general voice search traffic, as well as “local” traffic, up to 2% and 8% relative, respectively.

Full Paper

Bibliographic reference.  Chelba, Ciprian / Zhang, Xuedong / Hall, Keith (2015): "Geo-location for voice search language modeling", In INTERSPEECH-2015, 1438-1442.