DETECTING AND SUPPRESSING VOICE QUERIES

A computer-implemented method comprising receiving, at data processing hardware of a voice query processing server system, a plurality of voice query processing requests during a specified time interval, each voice query processing request comprising audio data and referencing a respective voice query. In response to receiving the plurality of voice query processing requests during the specified time interval and for each voice query processing request in at least a subset of the plurality of voice query processing requests received during the specified time interval, the data processing hardware processes the audio data of the respective voice query processing request to generate an acoustic fingerprint for the voice query referenced by the respective voice query processing request. The data processing hardware identifies a group of voice query processing requests that each reference a common voice query by determining matches among the acoustic fingerprints generated for each voice query processing request in at least the subset of the plurality of voice query processing requests. The data processing hardware suppresses performance of operations indicated by the common voice query.