Research Methods for the Analysis of Visual Emotion Cues in Animals: A Workshop Report

TL;DR

This workshop report discusses methods for analyzing visual emotion cues in animals using AI, highlighting challenges and recommendations for future research.

Contribution

The paper provides insights and recommendations from the first international workshop on animal emotion analysis using visual cues.

Findings

Workshop discussions identified challenges in defining and measuring animal emotions.

Key challenges include dataset collection and interpretation of visual cues.

Recommendations were provided to address these challenges based on expert discussions.

Abstract

The ways in which people and other animals express their emotions can vary significantly, making it difficult to identify and measure these emotions. Artificial intelligence (AI) has potential to aid people in this task, in a variety of settings, from research to animal care, but this technology introduces additional challenges which should be considered. With this in view, the first International Workshop on Research Methods for Animal Emotion Analysis (RM4AEA) was organised to present and debate some of the tools being used in this area, particularly to analyse visual cues of emotion in animals. The current work aims to report the results of this workshop, provide insights and recommendations for future work, and encourage further debate around this topic.

Animal emotions have been debated since Darwin’s pioneering work on the expression of such states in humans and other animals. Lately, interest in measuring animal emotions has been growing. Hence, a scientific discussion on this topic was needed, which translated into the organisation of the first International Workshop on Research Methods for Animal Emotion Analysis (RM4AEA). The current work aims to provide a concise yet critical examination of the current knowledge and methodological approaches discussed during the workshop, with a primary focus on visual behaviours. Although establishing conceptual definitions poses an initial challenge when studying animal emotions, the key challenges are found when collecting data sets, and when interpreting and analysing the information contained therein. This work also offers insights and recommendations to address these challenges, drawn from the closing panel discussion. By synthesising recent developments and expert perspectives shared during the workshop, this report aims to foster continued dialogue within the scientific community.

Figures