In-context Learning with Retrieved Demonstrations for Language Models: A Survey

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Luo, Man, Xu, Xin, Liu, Yue, Pasupat, Panupong, Kazemi, Mehran
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909147554906112
author Luo, Man
Xu, Xin
Liu, Yue
Pasupat, Panupong
Kazemi, Mehran
author_facet Luo, Man
Xu, Xin
Liu, Yue
Pasupat, Panupong
Kazemi, Mehran
contents Language models, especially pre-trained large language models, have showcased remarkable abilities as few-shot in-context learners (ICL), adept at adapting to new tasks with just a few demonstrations in the input context. However, the model's ability to perform ICL is sensitive to the choice of the few-shot demonstrations. Instead of using a fixed set of demonstrations, one recent development is to retrieve demonstrations tailored to each input query. The implementation of demonstration retrieval is relatively straightforward, leveraging existing databases and retrieval systems. This not only improves the efficiency and scalability of the learning process but also has been shown to reduce biases inherent in manual example selection. In light of the encouraging results and growing research in ICL with retrieved demonstrations, we conduct an extensive review of studies in this area. In this survey, we discuss and compare different design choices for retrieval models, retrieval training procedures, and inference algorithms.
format Preprint
id arxiv_https___arxiv_org_abs_2401_11624
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle In-context Learning with Retrieved Demonstrations for Language Models: A Survey
Luo, Man
Xu, Xin
Liu, Yue
Pasupat, Panupong
Kazemi, Mehran
Computation and Language
Artificial Intelligence
Information Retrieval
Language models, especially pre-trained large language models, have showcased remarkable abilities as few-shot in-context learners (ICL), adept at adapting to new tasks with just a few demonstrations in the input context. However, the model's ability to perform ICL is sensitive to the choice of the few-shot demonstrations. Instead of using a fixed set of demonstrations, one recent development is to retrieve demonstrations tailored to each input query. The implementation of demonstration retrieval is relatively straightforward, leveraging existing databases and retrieval systems. This not only improves the efficiency and scalability of the learning process but also has been shown to reduce biases inherent in manual example selection. In light of the encouraging results and growing research in ICL with retrieved demonstrations, we conduct an extensive review of studies in this area. In this survey, we discuss and compare different design choices for retrieval models, retrieval training procedures, and inference algorithms.
title In-context Learning with Retrieved Demonstrations for Language Models: A Survey
topic Computation and Language
Artificial Intelligence
Information Retrieval
url https://arxiv.org/abs/2401.11624