Zero-Shot Contextual Embeddings via Offline Synthetic Corpus Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lippmann, Philip, Yang, Jie
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911029877800960
author Lippmann, Philip
Yang, Jie
author_facet Lippmann, Philip
Yang, Jie
contents Context-aware embedding methods boost retrieval accuracy by conditioning on corpus statistics (e.g., term co-occurrence and topical patterns) extracted from neighboring documents. However, this context-aware approach requires access to the target corpus or requires domain-specific finetuning, posing practical barriers in privacy-sensitive or resource-constrained settings. We present ZEST, a zero-shot contextual adaptation framework that replaces real corpus access with a one-time offline synthesis of a compact proxy. Given only a handful exemplar documents representative of the general target domain, we use a multi-step hierarchical procedure to generate a synthetic context corpus of several hundred documents that aims to emulate key domain-specific distributions. At inference, the frozen context-aware encoder uses this proxy corpus -- without any finetuning or target corpus access -- to produce domain-adapted embeddings. Across the MTEB benchmark, ZEST's zero-shot synthetic context adaptation using only five example documents performs within 0.5% of models leveraging full target corpus access -- demonstrating remarkable efficacy without any retraining. ZEST thus provides a practical method for deploying high-performance, adaptable embeddings in constrained environments.
format Preprint
id arxiv_https___arxiv_org_abs_2506_23662
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Zero-Shot Contextual Embeddings via Offline Synthetic Corpus Generation
Lippmann, Philip
Yang, Jie
Computation and Language
Information Retrieval
Context-aware embedding methods boost retrieval accuracy by conditioning on corpus statistics (e.g., term co-occurrence and topical patterns) extracted from neighboring documents. However, this context-aware approach requires access to the target corpus or requires domain-specific finetuning, posing practical barriers in privacy-sensitive or resource-constrained settings. We present ZEST, a zero-shot contextual adaptation framework that replaces real corpus access with a one-time offline synthesis of a compact proxy. Given only a handful exemplar documents representative of the general target domain, we use a multi-step hierarchical procedure to generate a synthetic context corpus of several hundred documents that aims to emulate key domain-specific distributions. At inference, the frozen context-aware encoder uses this proxy corpus -- without any finetuning or target corpus access -- to produce domain-adapted embeddings. Across the MTEB benchmark, ZEST's zero-shot synthetic context adaptation using only five example documents performs within 0.5% of models leveraging full target corpus access -- demonstrating remarkable efficacy without any retraining. ZEST thus provides a practical method for deploying high-performance, adaptable embeddings in constrained environments.
title Zero-Shot Contextual Embeddings via Offline Synthetic Corpus Generation
topic Computation and Language
Information Retrieval
url https://arxiv.org/abs/2506.23662