Learned sparse retrieval
Learned sparse retrieval or sparse neural search is an approach to
vector embedding algorithms, and is claimed to perform better than either alone. The best-known sparse neural search systems are SPLADE[2] and its successor SPLADE v2.[3] Others include DeepCT,[4] uniCOIL,[5] EPIC,[6] DeepImpact,[7] TILDE and TILDEv2,[8] Sparta,[9] SPLADE-max, and DistilSPLADE-max.[3]
Some implementations of SPLADE have similar latency to Okapi BM25 lexical search while giving as good results as state-of-the-art neural rankers on in-domain data.[10]
The Official SPLADE model weights and training code is released under a Creative Commons NonCommercial license.[11] But there are other independent implementations of SPLADE++ (a variant of SPLADE models) that are released under permissive licenses.
SPRINT is a toolkit for evaluating neural sparse retrieval systems.[12]
External links
Notes
- S2CID 257585074.
- S2CID 235792467.
- ^ arXiv:2109.10086v1 [cs.IR].
- S2CID 218521094.
- arXiv:2106.14807 [cs.IR].
- S2CID 216641912.
- S2CID 233394068.
- arXiv:2108.08513 [cs.IR].
- arXiv:2009.13013 [cs.CL].
- S2CID 250340284.
- ^ "splade/LICENSE at main · naver/splade". GitHub. Retrieved 2023-08-25.
- S2CID 259949923.