ICICLE: Expanding Retrieval with In-Context Documents
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911722609049600 |
|---|---|
| author | Den, Yu-Chen Shih, Yung-Yu Tam, Zhi Rui Chen, Kuan-Yu Cheng, Pu-Jen Chen, Yun-Nung Yang, Eugene |
| author_facet | Den, Yu-Chen Shih, Yung-Yu Tam, Zhi Rui Chen, Kuan-Yu Cheng, Pu-Jen Chen, Yun-Nung Yang, Eugene |
| contents | Generative retrieval (GR) maps queries directly to document identifiers (docids) using parametric knowledge, However, this design makes corpus expansion costly: adding new documents requires updating model parameters to encode new document-docid associations incurs repeated training and catastrophic forgetting of previously indexed documents. In this work, we revisit incremental GR as an in-context retrieval problem, where newly added documents are supplied as inference-time document-docid evidence. We propose ICICLE, an in-context indexing framework that performs source-aware docid generation over both parametric memory and context-provided document-docid pairs. ICICLE combines a `[COPY]`-based routing mechanism, preference-based calibration, and large context adaptation to distinguish context-grounded retrieval from parametric retrieval. Experiments on MS MARCO and NQ320K show that ICICLE improves retrieval of newly introduced documents while preserving seen-document retention without corpus-specific retraining. Our analysis further shows that high-shot degradation is mainly caused by routing failure, highlighting source-selection calibration as a key bottleneck for scaling in-context generative retrieval. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_26902 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ICICLE: Expanding Retrieval with In-Context Documents Den, Yu-Chen Shih, Yung-Yu Tam, Zhi Rui Chen, Kuan-Yu Cheng, Pu-Jen Chen, Yun-Nung Yang, Eugene Information Retrieval Artificial Intelligence Generative retrieval (GR) maps queries directly to document identifiers (docids) using parametric knowledge, However, this design makes corpus expansion costly: adding new documents requires updating model parameters to encode new document-docid associations incurs repeated training and catastrophic forgetting of previously indexed documents. In this work, we revisit incremental GR as an in-context retrieval problem, where newly added documents are supplied as inference-time document-docid evidence. We propose ICICLE, an in-context indexing framework that performs source-aware docid generation over both parametric memory and context-provided document-docid pairs. ICICLE combines a `[COPY]`-based routing mechanism, preference-based calibration, and large context adaptation to distinguish context-grounded retrieval from parametric retrieval. Experiments on MS MARCO and NQ320K show that ICICLE improves retrieval of newly introduced documents while preserving seen-document retention without corpus-specific retraining. Our analysis further shows that high-shot degradation is mainly caused by routing failure, highlighting source-selection calibration as a key bottleneck for scaling in-context generative retrieval. |
| title | ICICLE: Expanding Retrieval with In-Context Documents |
| topic | Information Retrieval Artificial Intelligence |
| url | https://arxiv.org/abs/2605.26902 |