Standard Metrics for Assessing Human-Machine Team Performance

EasyChair Preprint no. 14091

Abstract

The integration of artificial intelligence (AI) and machine learning into various domains has fostered the emergence of human-machine teams, where humans and machines collaborate to achieve common goals. This symbiotic relationship necessitates the development of robust metrics to evaluate the performance and efficacy of these teams. This paper reviews the existing literature on human-machine team performance assessment and proposes a comprehensive framework for standard metrics. Key metrics include task completion time, accuracy, reliability, adaptability, user satisfaction, and the cognitive load on human team members. We also explore advanced metrics such as team synergy, decision-making quality, and the effectiveness of communication between human and machine agents. Additionally, the paper addresses the importance of contextual factors, such as the complexity of tasks and the operational environment, which significantly influence performance outcomes. By establishing a standardized set of metrics, this research aims to provide a foundation for systematic evaluation, facilitate comparative studies, and guide the design and optimization of future human-machine teaming systems.

Keyphrases: Data collection techniques, Human Factors, human-machine teaming, performance metrics, Task Performance, team interaction

@Booklet{EasyChair:14091,
  author = {John Owen},
  title = {Standard Metrics for Assessing Human-Machine Team Performance},
  howpublished = {EasyChair Preprint no. 14091},

  year = {EasyChair, 2024}}