Sub-SA: Strengthen In-context Learning via Submodular Selective Annotation

TL;DR

Sub-SA introduces a submodular function-based method to select in-context examples efficiently, reducing annotation costs and enhancing performance in large language models through a greedy selection algorithm.

Contribution

It proposes a novel submodule-based selective annotation method using submodular functions and RPR to improve in-context learning with lower annotation costs.

Findings

Effective subset selection for annotation via submodular functions.

RPR balances diversity and representativeness in data selection.

Greedy search algorithm efficiently selects high-quality in-context examples.

Abstract

In-context learning (ICL) leverages in-context examples as prompts for the predictions of Large Language Models (LLMs). These prompts play a crucial role in achieving strong performance. However, the selection of suitable prompts from a large pool of labeled examples often entails significant annotation costs. To address this challenge, we propose Sub-SA (Submodular Selective Annotation), a submodule-based selective annotation method. The aim of Sub-SA is to reduce annotation costs while improving the quality of in-context examples and minimizing the time consumption of the selection process. In Sub-SA, we design a submodular function that facilitates effective subset selection for annotation and demonstrates the characteristics of monotonically and submodularity from the theoretical perspective. Specifically, we propose RPR (Reward and Penalty Regularization) to better balance the diversity and representativeness of the unlabeled dataset attributed to a reward term and a penalty term, respectively. Consequently, the selection for annotations can be effectively addressed with a simple yet effective greedy search algorithm based on the submodular function. Finally, we apply the similarity prompt retrieval to get the examples for ICL.