Replacing standard self-attention with performer attention in the pooling layer of a language identification model improves average accuracy on VoxPopuli, FLEURS, and VoxLingua, while agent attention is comparable and theoretically cheaper.
Language identifica- tion from short segments of speech,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
On the use of Performer and Agent Attention for Spoken Language Identification
Replacing standard self-attention with performer attention in the pooling layer of a language identification model improves average accuracy on VoxPopuli, FLEURS, and VoxLingua, while agent attention is comparable and theoretically cheaper.