Decision making in open agent systems

A Eck, LK Soh, P Doshi - AI Magazine, 2023 - Wiley Online Library
In many real‐world applications of AI, the set of actors and tasks are not constant, but
instead change over time. Robots tasked with suppressing wildfires eventually run out of …

Efficient Human-AI Coordination via Preparatory Language-based Convention

C Guan, L Zhang, C Fan, Y Li, F Chen, L Li… - ar** intelligent agents capable of seamless coordination with humans is a critical
step towards achieving artificial general intelligence. Existing methods for human-AI …

Mutual theory of mind in human-ai collaboration: An empirical study with llm-driven ai agents in a real-time shared workspace task

S Zhang, X Wang, W Zhang, Y Chen, L Gao… - arxiv preprint arxiv …, 2024 - arxiv.org
Theory of Mind (ToM) significantly impacts human collaboration and communication as a
crucial capability to understand others. When AI agents with ToM capability collaborate with …

[PDF][PDF] Multi-objective Optimization-based Selection for Quality-Diversity by Non-surrounded-dominated Sorting.

RJ Wang, K Xue, H Shang, C Qian, H Fu, Q Fu - IJCAI, 2023 - ijcai.org
Abstract Quality-Diversity (QD) algorithms, a subset of evolutionary algorithms, maintain an
archive (ie, a set of solutions) and simulate the natural evolution process through iterative …

Open Human-Robot Collaboration Systems (OHRCS): A Research Perspective

PS Suresh, D Romeres, P Doshi… - 2024 IEEE 6th …, 2024 - ieeexplore.ieee.org
Human-robot collaboration (HRC) is the paradigm of humans and robots working
synergistically in a shared workspace toward common goals. Prior research models such …

Fast peer adaptation with context-aware exploration

L Ma, Y Wang, F Zhong, SC Zhu, Y Wang - arxiv preprint arxiv …, 2024 - arxiv.org
Fast adapting to unknown peers (partners or opponents) with different strategies is a key
challenge in multi-agent games. To do so, it is crucial for the agent to probe and identify the …

Few is More: Task-Efficient Skill-Discovery for Multi-Task Offline Multi-Agent Reinforcement Learning

X Wang, Z Li, H Zhong, L Huang - arxiv preprint arxiv:2502.08985, 2025 - arxiv.org
As a data-driven approach, offline MARL learns superior policies solely from offline datasets,
ideal for domains rich in historical data but with high interaction costs and risks. However …

Beyond Single Stationary Policies: Meta-Task Players as Naturally Superior Collaborators

H Wang, Z Tian, Y Song, X Zhang, Z Cai - The Thirty-eighth Annual … - openreview.net
In human-AI collaborative tasks, the distribution of human behavior, influenced by mental
models, is non-stationary, manifesting in various levels of initiative and different …

Opponent Transformer: Modeling Opponent Policies as a Sequence Problem

C Wallace, U Siddique, Y Cao - Coordination and Cooperation for Multi … - openreview.net
The ability of an agent to understand the intentions of others in a multi-agent system, also
called opponent modeling, is critical for the design of effective local control policies. One …