activity
20192026
most citedDistill-VQ: Learning Retrieval Oriented Vector Quantization By Distilling Knowledge from Dense Embeddings

4 citations · 5 across the 10 of their papers we have counts for

collaborators
Showing 2019Show all

1 paper · 1 filter

cs.IR2019

Learning Fast Matching Models from Weak Annotations

Xue Li, Zhipeng Luo, Hao Sun +5

This paper proposes a novel training scheme for fast matching models in Search Ads, which is motivated by the real challenges in model training. The first challenge stems from the…