More Like this

1. Simple Temporal Networks for Improvisational Teamwork

   Morgan, M.; Schalkwyk, J.; Wang, H.; Davalos, H.; Martinez, R.; Rohilla, V.; Boerkoel, J. (March 2022, AAAI Spring Symposium Series on Can We Talk? How to Design Multi-Agent Systems In the Absence of Reliable Communications)

   When communication between teammates is limited to observations of each other’s actions, agents may need to improvise to stay coordinated. Unfortunately, current methods inadequately capture the uncertainty introduced by a lack of direct communication. This paper augments existing frameworks to introduce Simple Temporal Networks for Improvisational Teamwork (STN-IT) — a formulation that captures both the temporal dependencies and uncertainties between agents who need to coordinate, but lack reliable communication. We define the notion of strong controllability for STN-ITs, which establishes a static scheduling strategy for controllable agents that produces a consistent team schedule, as long as non-communicative teammates act within known problem constraints. We provide both an exact and approximate approach for finding strongly controllable schedules, empirically demonstrate the trade-offs between these two approaches on a benchmark of STN-ITs, and show analytically that the exact method is correct. In addition, we provide an empirical analysis of the exact and approximate approaches’ efficiency 

   more » « less
2. An Integrated Architecture for Common Ground in Collaboration

   Geib, Christopher; George, Denson; Khalid, Baber; Magnotti, Richard; Stone, Matthew (November 2022, http://www.cogsys.org/)

   Effective teamwork depends on teammates’ ability to maintain common ground: mutual knowledge about the relevant state of the world and the relevant status of teammates’ actions and plans. This ability integrates diverse skills of reasoning and communication: agents can track common ground by recognizing and registering public updates to ongoing activity, but when this evidence is incomplete, agents may need to describe what they are doing or ask what others are doing. In this paper, we introduce an architecture for integrating these diverse skills to maintain common ground in human–AI teamwork. Our approach offers unique advantages of simplicity, modularity, and extensibility by leveraging generic tools for plan recognition, planning, natural language understanding and generation, and dialogue management. Worked examples illustrate how linguistic and practical reasoning complement each other in the realization of key interactive skills. 

   more » « less
3. Effect of Adaptive Communication Support on Human-AI Collaboration

   Liu, Shipeng; Shrutika, FNU; Zhang, Boshen; Huang, Zhehui; Qian, Feifei (November 2024, AAAI 2025 Workshop on Advancing LLM-Based Multi-Agent Collaboration)

   Effective human-AI collaboration requires agents to adopt their roles and levels of support based on human needs, task requirements, and complexity. Traditional human-AI teaming often relies on a pre-determined robot communication scheme, restricting teamwork adaptability in complex tasks. Leveraging the strong communication capabilities of Large Language Models (LLMs), we propose a Human-Robot Teaming Framework with Multi-Modal Language feedback (HRT-ML), a framework designed to enhance human-robot interaction by adjusting the frequency and content of language-based feedback. The HRT-ML framework includes two core modules: a Coordinator for high-level, low-frequency strategic guidance and a Manager for task-specific, high-frequency instructions, enabling passive and active interactions with human teammates. To assess the impact of language feedback in collaborative scenarios, we conducted experiments in an enhanced Overcooked-AI game environment with varying levels of task complexity (easy, medium, hard) and feedback frequency (inactive, passive, active, superactive). Our results show that as task complexity increases relative to human capabilities, human teammates exhibited stronger preferences toward robotic agents that can offer frequent, proactive support. However, when task complexities exceed the LLM's capacity, noisy and inaccurate feedback from superactive agents can instead hinder team performance, as it requires human teammates to increase their effort to interpret and respond to the large amount of communications, with limited performance return. Our results offer a general principle for robotic agents to dynamically adjust their levels and frequencies of communication to work seamlessly with humans and achieve improved teaming performance. 

   more » « less
4. New Perspectives on Flexibility in Simple Temporal Planning

   https://doi.org/10.1609/icaps.v28i1.13907  

   Huang, Amy; Lloyd, Liam; Omar, Mohamed; Boerkoel, James (June 2018, Proceedings of the International Conference on Automated Planning and Scheduling)

   de_Weerdt, Mathijs; Koenig, Sven; Röger, Gabriele; Spaan, Matthijs (Ed.)

   Flexibility is generally agreed to be a desirable feature of a Simple Temporal Network (STN). However, exactly what flexibility attempts to measure has varied, making it difficult to objectively evaluate flexibility metrics. Further, past metrics tend to lose information or exhibit other undesirable properties when aggregating the flexibility measures of individual events across an entire STN. Our work is driven by the realization that the solution space of an STN is a convex polyhedron whose geometric properties convey useful information about the STN. These geometric inspirations lead to measures of an STN solution space and also motivate a set of desiderata for general flexibility metrics. We also put forth two new geometrically-inspired flexibility metrics that have some theoretical advantages over existing metrics. Finally, we comprehensively evaluate both new and existing flexibility metrics against our proposed desiderata. 

   more » « less
5. DREAM: An Algorithm for Mitigating the Overhead of Robust Rescheduling

   https://doi.org/10.1609/icaps.v29i1.3454  

   Abrahams, Jordan R; Chu, David A; Diehl, Grace; Knittel, Marina; Lin, Judy; Lloyd, William; Boerkoel_Jr, James C; Jeremy, Frank (July 2019, Proceedings of the International Conference on Automated Planning and Scheduling)

   Benton, J; Lipovetzky, Nir; Onaindia, Eva; Smith, David E; Srivastava, Siddharth (Ed.)

   Generating and executing temporal plans is difficult in uncertain environments. The current state-of-the-art algorithm for probabilistic temporal networks maintains a high success rate by rescheduling frequently as uncertain events are resolved, but this approach involves substantial resource overhead due to computing and communicating new schedules between agents. Aggressive rescheduling could thus reduce overall mission duration or success in situations where agents have limited energy or computing power, and may not be feasible when communication is limited. In this paper, we propose new approaches for heuristically deciding when rescheduling is most efficacious. We propose two compatible approaches, Allowable Risk and Sufficient Change, that can be employed in combination to compromise between the computation rate, the communication rate, and success rate for new schedules. We show empirically that both approaches allow us to gracefully trade success rate for lower computation and/or communication as compared to the state-of-the-art dynamic scheduling algorithm. 

   more » « less