Once is Enough: A Light-Weight Cross-Attention for Fast Sentence Pair Modeling
This addresses efficiency issues for NLP practitioners working on tasks like answer selection and NLI, offering a significant speedup with minimal performance loss, though it is an incremental improvement over existing fast architectures.
The paper tackles the high computational cost of cross-attention in transformer models for sentence pair tasks by introducing MixEncoder, a light-weight cross-attention mechanism that speeds up processing by over 113x while maintaining comparable performance.
Transformer-based models have achieved great success on sentence pair modeling tasks, such as answer selection and natural language inference (NLI). These models generally perform cross-attention over input pairs, leading to prohibitive computational costs. Recent studies propose dual-encoder and late interaction architectures for faster computation. However, the balance between the expressive of cross-attention and computation speedup still needs better coordinated. To this end, this paper introduces a novel paradigm MixEncoder for efficient sentence pair modeling. MixEncoder involves a light-weight cross-attention mechanism. It conducts query encoding only once while modeling the query-candidate interaction in parallel. Extensive experiments conducted on four tasks demonstrate that our MixEncoder can speed up sentence pairing by over 113x while achieving comparable performance as the more expensive cross-attention models.