Enhancing Social Robots through Resilient AI

TL;DR

This paper investigates resilience as a core requirement for social robots operating in sensitive contexts, notably elderly care. It defines resilience using absorptive, adaptive, and transformative capacities and argues for embedding resilience from data preparation (data augmentation) to model selection (robust, stable models) in FER and NLP. The authors review a corpus of 20 articles published between 2021 and 2025 to identify effective strategies for reliability and adaptability in noisy, uncertain environments. The RAISE project is introduced as a focused program to apply resilient AI to elder care, with this review laying groundwork for safer, more autonomous, and socially capable robots.

Abstract

As artificial intelligence continues to advance and becomes more integrated into sensitive areas like healthcare, education, and everyday life, it's crucial for these systems to be both resilient and robust. This paper shows how resilience is a fundamental characteristic of social robots, which, through it, ensure trust in the robot itself-an essential element especially when operating in contexts with elderly people, who often have low trust in these systems. Resilience is therefore the ability to operate under adverse or stressful conditions, even when degraded or weakened, while maintaining essential operational capabilities.