1 paper · 1 filter
Huang Xie, Khazar Khorrami, Okko Räsänen +1
Audio-text relevance learning refers to learning the shared semantic properties of audio samples and textual descriptions. The standard approach uses binary relevances derived from…