Researchers have introduced Passage-Driven Multi-ID Retrieval (PDMR), a new generative retrieval framework that enhances document retrieval by assigning multiple passage-level identifiers to each document. This approach allows models to align queries with specific semantic facets of a document, moving beyond the limitations of single-identifier systems that can lead to lossy representations. PDMR has demonstrated improved performance on benchmark datasets like NQ320K and MS MARCO Document, outperforming existing generative and non-generative methods in recall and mean reciprocal rank. AI
IMPACT This new retrieval method could improve the accuracy and robustness of search systems by better handling multi-faceted documents.
RANK_REASON The cluster contains a research paper detailing a new method for information retrieval. [lever_c_demoted from research: ic=1 ai=1.0]
Read on arXiv cs.IR (Information Retrieval) →
- alphaXiv
- CatalyzeX Code Finder for Papers
- CORE Recommender
- DagsHub
- Generative Retrieval
- Gotit.pub
- Hugging Face
- MS MARCO Document
- NQ320K
- Passage-Driven Multi-ID Retrieval
- PDMR
- ScienceCast
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →