For learned sparse retrieval, an encoder-decoder model that copies input to the decoder and max-pools MLM logits over all positions outperforms encoder-only and decoder-only backbones on MS MARCO.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Leveraging Decoder Architectures for Learned Sparse Retrieval
For learned sparse retrieval, an encoder-decoder model that copies input to the decoder and max-pools MLM logits over all positions outperforms encoder-only and decoder-only backbones on MS MARCO.