Search-based Optimisation of LLM Learning Shots for Story Point Estimation

TL;DR

This paper uses available Search-Based methods to optimise the number and combination of examples that can improve an LLM's estimation performance, when it is used to estimate story points for new agile tasks.

Abstract

One of the ways Large Language Models (LLMs) are used to perform machine learning tasks is to provide them with a few examples before asking them to produce a prediction. This is a meta-learning process known as few-shot learning. In this paper, we use available Search-Based methods to optimise the number and combination of examples that can improve an LLM's estimation performance, when it is used to estimate story points for new agile tasks. Our preliminary results show that our SBSE technique improves the estimation performance of the LLM by 59.34% on average (in terms of mean absolute error of the estimation) over three datasets against a zero-shot setting.

Figures (2)

• Figure 1: The prompt used with the LLM to estimate story points for a target issue, in a zero/few-shot learning (depending on the length of the example issues list).
• Figure 2: Pareto fronts achieved by our proposed approach for three projects.