The Socio-Cognitive Engineering method fosters the specification and maintenance of the design rationale, including the abstraction of so-called design patterns: generic, re-usable solutions to recurring design problems. They have been and are being developed in numerous research and development projects, available for application in other projects and for other development teams. It should be noted that design patterns have been proposed for different kinds of design problems and application domains (i.e., originated from architecture designs and taken up in domains like software programming, workflow engineering and interaction design). We distinguish Team Design Patterns (TDP) and Interaction Design Patterns (IDP).
Note that current research is aiming to improve the pattern engineering methods and practices for the development of resonsible and effective human-machine collaboration. In a Dutch research programme on Hybrid Intelligence, a specific project is developing, applying and testing a design pattern language for Hybrid Intelligence, see https://ii.tudelft.nl/HIDesignPatterns. See also:
- Tajaddini, M., Brinkman, W. P., ten Teije, A., & Neerincx, M. A. (2021). A Design Pattern Language for Hybrid Intelligent Teams. In 33rd Benelux Conference on Artificial Intelligence and30th Belgian-Dutch Conference on Machine Learning (pp. 723-725).
The SCE method will include the new insights, models and techniques when appropriate. Furthermore, we will make the first version of the design pattern library available in WiSCE before the summer of 2022 (see also the workshop on Human-Centered Design of Symbiotic Hyrbid Intelligence, https://ii.tudelft.nl/humancenteredsymbioticHI , which will provide HI-patterns of current research in this area) .
Team Design Patterns (TDP)
Team Design Pattners (TDPs) are applied to the problem of developing Human-Agent/Robot-Teams, and can be defined as a description of generic reusable behaviors of actors of effective and resilient teamwork.
Hybrid (human-AI) disease management
Van Stijn et al. (2021) present Team Design Patterns for mitigating harmful biases in machine learning algorithms, based on the research and development of hybrid intelligence (HI) system for diabetes management. It was shown that TDPs are a useful method to unambiguously describe solutions for diverse HI design problems with a moral component on varying abstraction levels, usable by a heterogeneous group of multidisciplinary researchers.
- Van Stijn, J.J., Neerincx, M.A., ten Teije, A.T., and Vethman, S. (2021). Team Design Patterns for Moral Decisions in Hybrid Intelligent Systems: A Case Study of Bias Mitigation. AAAI 2021 Spring Symposium on Combining Machine Learning and Knowledge Engineering (AAAI-MAKE 2021)
Allocation of tasks and responsibilities in HART
Van der Waa et al. (2020) provide some high-level patterns for the allocation of moral decision-making in human-agent/robot teams (HART):
- van der Waa, J., van Diggelen, J., Siebert, L. C., Neerincx, M., & Jonker, C. (2020). Allocation of Moral Decision-Making in Human-Agent Teams: A Pattern Approach. In International Conference on Human-Computer Interaction (pp. 203-220). Springer, Cham.
Diggelen et al. (2018) provide a team design pattern language that can be used for the development of artificial team members (e.g. robots, avatars). These patterns can be constructive or destructive, applying to an individual team-member or the overall for team. The patterns are defined with reference to a team ontology. Example from the Space and Railway domains provide some example patterns (e.g. for local and global problem solving via a huddle, and for after-action team reflection via a Resiliencer display).
- van Diggelen, J., Neerincx, M., Peeters, M., & Schraagen, J. M. (2018). Developing effective and resilient human-agent teamwork using team design patterns. IEEE intelligent systems, 34(2), 15-24.
Robot-assisted disaster response teams
The design patterns of Neerincx et al. (2016) for this domain, did not (yet) make a distinction between Team and Interaction Design Patterns, there descriptions provide a mix of them. The Design Patterns address the mutual and context-dependent behaviors of the human and agent/robot team-members during disaster response missions. The pattern descriptions capture four key concepts: The Actor can be Human, Agent or Robot, the Relationship between actors can be Supervisory and/or Collaborative, actors can perform their work at the Same (co-location) or a Distant (distributed) Location, and the Pattern status can be Proto (i.e., in construction) or Grounded (e.g., empirically validated in an experiment). Three example patterns are provided. The first example centers on obtaining adjustable working agreements between humans and agents to establish flexible and adaptive teamwork, supporting dynamic and adaptive human-agent (sub)task allocation. For specific work contexts, the human can set agreements with the agent on how the tasks will be allocated. The second design pattern centers on the demand for an operator to stay in vicinity of the workstation when an event or situation may appear that will ask for immediate action. The third pattern focuses on the management of multiple interactions between the human and agent. The patterns are presented in:
- Neerincx, M. A., van Diggelen, J., & van Breda, L. (2016). Interaction design patterns for adaptive human-agent-robot teamwork in high-risk domains. In International conference on engineering psychology and cognitive ergonomics (pp. 211-220). Springer, Cham.
Van Zoelen et al. (2021) are developing a method for the generation of "human-robot co-learning patterns" out of succesfull recurring behaviors in joint human-robot performances:
Van Zoelen, E. M., Van Den Bosch, K., & Neerincx, M. (2021). Becoming team members: identifying interaction patterns of mutual adaptation for human-robot co-learning. Frontiers in Robotics and AI, 8.
Interaction Design Patterns (IDP)
eXplainable Artificial Intelligence (XAI)
The following paper presents how to integrate human factors into the development processes of AI-generated explanations in a mental health use case, providing a set of generic interaction design patterns for the explanations that AI-agents can provide:.
Schoonderwoerd, T. A., Jorritsma, W., Neerincx, M. A., & Van Den Bosch, K. (2021). Human-centered XAI: Developing design patterns for explanations of clinical decision support systems. International Journal of Human-Computer Studies, 154, 102684.
Robot-assisted disaster response teams
Note that Neerincx et al. (2016) provide a mixture of Team and Interaction Design Patterns as described above. Further, Mioch et al. (2014) provide IDPs for a team-awareness display, which has been tested with teams of fire fighters, ground and aerial robots, in several realistic earth quake scenarios:
- Mioch, T., Ledegang, W., Paulissen, R., Neerincx, M. A., & van Diggelen, J. (2014). Interaction design patterns for coherent and re-usable shape specifications of human-robot collaboration. In Proceedings of the 2014 ACM SIGCHI symposium on Engineering interactive computing systems (pp. 75-83).
Child-robot interaction in the health domain
Ligthart et a. (2000) provide three interaction design patterns for an interactive storytelling robot. In the patterns, children can adjust the story by talking with the robot, reenacting parts of the story with the robot, and recording self-made sound effects. These patterns prove to support children’s engagement and agency.
- Ligthart, M. E., Neerincx, M. A., & Hindriks, K. V. (2020). Design patterns for an interactive storytelling robot to support children's engagement and agency. In Proceedings of the 2020 ACM/IEEE International Conference on Human-Robot Interaction (pp. 409-418).