Semi-Parametric Retrieval via Binary Bag-of-Tokens Index
This addresses indexing efficiency and cost issues for information retrieval applications, offering a novel hybrid approach that is not purely incremental but combines neural and term-based methods.
The paper tackles the problem of efficient and cost-effective indexing for neural retrieval systems by introducing SiDR, a bi-encoder framework that decouples retrieval index from neural parameters, achieving BM25-like indexing complexity with better effectiveness across 16 benchmarks, such as outperforming BM25 on all in-domain datasets.
Information retrieval has transitioned from standalone systems into essential components across broader applications, with indexing efficiency, cost-effectiveness, and freshness becoming increasingly critical yet often overlooked. In this paper, we introduce SemI-parametric Disentangled Retrieval (SiDR), a bi-encoder retrieval framework that decouples retrieval index from neural parameters to enable efficient, low-cost, and parameter-agnostic indexing for emerging use cases. Specifically, in addition to using embeddings as indexes like existing neural retrieval methods, SiDR supports a non-parametric tokenization index for search, achieving BM25-like indexing complexity with significantly better effectiveness. Our comprehensive evaluation across 16 retrieval benchmarks demonstrates that SiDR outperforms both neural and term-based retrieval baselines under the same indexing workload: (i) When using an embedding-based index, SiDR exceeds the performance of conventional neural retrievers while maintaining similar training complexity; (ii) When using a tokenization-based index, SiDR drastically reduces indexing cost and time, matching the complexity of traditional term-based retrieval, while consistently outperforming BM25 on all in-domain datasets; (iii) Additionally, we introduce a late parametric mechanism that matches BM25 index preparation time while outperforming other neural retrieval baselines in effectiveness.