Cross-disciplinary concept borrowing : a critical revision of a robustness analysis for automatic emotion recognition in conversations

Abstract

Affective computing is an inherently interdisciplinary field, drawing knowledge from diverse disciplines such as psychology, neuroscience, sociology, and cultural studies. Typically, this knowledge is adopted pragmatically and action-oriented to inform research in the field. This paper critically examines a robustness analysis in affective computing to explore the conceptual borrowing of terms, theories, and methods from these reference disciplines. It highlights the challenges when knowledge from fields with differing foundations - such as the humanities and social sciences - is integrated into computer science. The paper argues that these disciplinary differences, especially in terms of research practices and theoretical frameworks, can lead to misunderstandings and simplification during the appropriation process, potentially limiting the effectiveness and accuracy of affective computing systems.