Human-machine Symbiosis: A Multivariate Perspective for Physically Coupled Human-machine Systems

Jairo Inga, Miriam Ruess, Jan Heinrich Robens, Thomas Nelius, Sean Kille, Philipp Dahlinger, Roland Thomaschke, Gerhard Neumann, Sven Matthiesen, Sören Hohmann, Andrea KieselInt. J. Hum. Comput. Stud 2021 · OpenReview

TL;DR

This paper proposes a new multivariate framework for understanding human-machine symbiosis, moving beyond single-dimension views. By considering task, interaction, performance, and experience, the authors offer a more holistic approach to designing and evaluating physically coupled human-machine systems, aiming for truly collaborative and beneficial partnerships.

What Problem Does This Paper Address?

The concept of "symbiosis" is increasingly used to describe advanced human-machine interaction, particularly in physically coupled systems where humans and machines work closely together. However, there's a lack of a unified understanding of what truly constitutes symbiosis. Existing frameworks often focus on single aspects like shared control or collaboration, failing to capture the multifaceted nature of a symbiotic relationship. This ambiguity hinders the development of systems that genuinely leverage the strengths of both humans and machines for mutual benefit, leading to a need for a more comprehensive and interdisciplinary perspective.

Methods

The authors propose a multivariate perspective for human-machine symbiosis, breaking it down into four key dimensions:

  • Task Dimension: This dimension analyzes how complex tasks are decomposed and managed across different levels of abstraction (decomposition, decision, and action). It considers how humans and machines can collaborate at each of these levels.
  • Interaction Dimension: This focuses on the mechanisms of interplay between humans and machines. Key features include:
  • Highly skilled and coordinated joint action.
  • Flexible and context-sensitive shifting of automation levels.
  • Congruent individual representations of intentions and expectations.
  • Continuous feedback loops within a perception-action cycle.
  • Performance Dimension: This dimension evaluates the overall effectiveness of the symbiotic system. It goes beyond individual capabilities, aiming for synergistic effects where the combined performance is greater than the sum of its parts. Evaluation metrics include efficiency, error rates, and quality of work.
  • Experience Dimension: This dimension specifically addresses the human user's subjective experience. It encompasses:
  • Flow Experience: The feeling of deep immersion and engagement when task challenges align with perceived skills.
  • Acceptance: The user's willingness to adopt and integrate the technology.
  • Sense of Agency: The user's feeling of control over their actions and their consequences.
  • Embodiment: The feeling of the machine being integrated into the user's body scheme.

The paper emphasizes an interdisciplinary approach, drawing from engineering, computer science, and psychology.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Symbiosis as a Spectrum: The paper implicitly suggests that symbiosis isn't an all-or-nothing state but rather a high form of interaction that can be achieved through careful design across multiple dimensions.
  • Beyond Shared Control: While shared control is a foundation, symbiosis requires a deeper level of mutual understanding and adaptation, moving towards a "true collaboration" where partners are considered equals.
  • The Importance of User Experience: The inclusion of the "Experience Dimension" is crucial. It highlights that for a system to be truly symbiotic, it must not only perform well but also be a positive and engaging experience for the human user.
  • Interplay Between Dimensions: The framework emphasizes that these four dimensions are interconnected. For instance, effective interaction (Interaction Dimension) leads to better performance (Performance Dimension), which in turn can enhance the user's experience (Experience Dimension).
  • Flexibility in Automation: Symbiotic systems should allow for dynamic and context-sensitive shifts in automation levels, enabling either the human or the machine to take the lead as appropriate for the task.

Key Results

The paper is primarily a conceptual framework and does not present empirical results or quantitative performance metrics. Its key contribution lies in proposing the multivariate perspective itself. The authors conjecture that this approach will lead to:

  • Higher overall performance: Synergistic effects beyond individual capabilities.
  • Increased efficiency: Optimized work and time utilization.
  • Reduced errors: Due to better mutual understanding and adaptation.
  • Improved user experience: Including flow, acceptance, sense of agency, and embodiment.

Limitations and Future Work

The authors acknowledge that while the multivariate perspective is a framework, its detailed specification requires further work. Specific areas for future research include:

  • Task Dimension Refinement: Amplifying the task dimension to include more granular levels like operational or decomposition levels.
  • Interaction Sophistication: Developing more sophisticated feedback control loops and recursive intention recognition for real-time updating of representations.
  • Performance Specification: Detailing the "intention recognition" and "complementary strengths" aspects of performance.
  • Experience Dimension Expansion: Further exploring the interplay of flow, acceptance, sense of agency, and embodiment, and potentially incorporating other factors like trust and ethical considerations.
  • Trade-offs: Investigating optimal trade-offs between the different dimensions, as singular optimization in one dimension might not lead to the best overall outcome.

Summary updated

Abstract

The notion of symbiosis has been increasingly mentioned in research on physically coupled human-machine systems. Yet, a uniform specification on which aspects constitute human-machine symbiosis is missing. By combining the expertise of different disciplines, we elaborate on a multivariate perspective of symbiosis as the highest form of physically coupled human-machine systems. Four dimensions are considered: Task, interaction, performance, and experience. First, human and machine work together to accomplish a common task conceptualized on both a decision and an action level (task dimension). Second, each partner possesses an internal representation of own as well as the other partner's intentions and influence on the environment. This alignment, which is the core of the interaction, constitutes the symbiotic understanding between both partners, being the basis of a joint, highly coordinated and effective action (interaction dimension). Third, the symbiotic interaction leads to synergetic effects regarding the intention recognition and complementary strengths of the partners, resulting in a higher overall performance (performance dimension). Fourth, symbiotic systems specifically change the user's experiences, like flow, acceptance, sense of agency, and embodiment (experience dimension). This multivariate perspective is flexible and generic and is also applicable in diverse human-machine scenarios, helping to bridge barriers between different disciplines.
BibTeX
@article{inga2021human,
  author = {Jairo Inga and Miriam Ruess and Jan Heinrich Robens and Thomas Nelius and Sean Kille and Philipp Dahlinger and Roland Thomaschke and Gerhard Neumann and Sven Matthiesen and Sören Hohmann and Andrea Kiesel},
  title = {{Human-machine Symbiosis: A Multivariate Perspective for Physically Coupled Human-machine Systems}},
  journal = {Int. J. Hum. Comput. Stud.},
  year = {2021},
  url = {https://openreview.net/forum?id=7PjkpPQ0hS},
}
Figure