Can Information Behaviour Inform Machine Learning?

TL;DR

This paper discusses how human information behaviour research can enhance machine learning by providing insights into information needs, context, bias, and misinformation, especially in the development of foundation models.

Contribution

It highlights the potential for integrating human information behaviour insights into machine learning to address challenges like bias and context understanding.

Findings

Human information behaviour can inform machine learning models.

Insights into bias and misinformation can improve model fairness.

Understanding context enhances model relevance and accuracy.

Abstract

The objective of this paper is to explore the opportunities for human information behaviour research to inform and influence the field of machine learning and the resulting machine information behaviour. Using the development of foundation models in machine learning as an example, the paper illustrates how human information behaviour research can bring to machine learning a more nuanced view of information and informing, a better understanding of information need and how that affects the communication among people and systems, guidance on the nature of context and how to operationalize that in models and systems, and insights into bias, misinformation, and marginalization. Despite their clear differences, the fields of information behaviour and machine learning share many common objectives, paradigms, and key research questions. The example of foundation models illustrates that human information behaviour research has much to offer in addressing some of the challenges emerging in the nascent area of machine information behaviour.