Research map / Multi-agent systems and games / Human-AI and goal-conditioned
AI-Assisted Human Teamwork
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Doctoral research proposal for an automated task-time intervention system that detects shared-mental-model misalignment in human teams via learned intent-aware teamwork models (BTIL) and issues interventions (TIC) to improve teamwork.
Problem
Teams in time-critical collaborative tasks suffer from partial observability, costly communication, and difficulty estimating teammates' states and intents, leading to preventable errors and poor performance; expert-crafted teamwork rules are costly to build.
Approach
Learn intent-aware team policies and intent dynamics from past demonstrations (Bayesian Team Imitation Learner, BTIL), use them to detect imperfect coordination via intent compatibility, and deliver cost-benefit-balanced task-time interventions (TIC) that inform the team of compatible intents.
Key results
- Proposes BTIL, a Bayesian multi-agent imitation learner that learns intent-aware team policies and intent dynamics from demonstrations using mean-field variational inference, claimed to be sample- and label-efficient.
- Proposes TIC, a task-time intervention system that balances intervention costs and benefits using heuristic strategies over inferred intent distributions.
- Reports that human demonstrations have been collected under IRB approval and a validation experiment protocol was recently approved; no quantitative results are reported.
Limitations
- BTIL is Bayesian and struggles to scale to high-dimensional or continuous state spaces.
- Solving the TIC objective is computationally intractable, so heuristic strategies are used.
- TIC assumes the set of possible intents is known a priori, which is often unrealistic.
Experimental setup
Datasets: human demonstrations of teamwork collected under IRB approval
Builds on: BTIL Option-GAIL generative adversarial imitation learning
Closest papers by meaning
- Designs for Enabling Collaboration in Human-Machine Teaming via Interactive and Explainable Systems (NeurIPS 2024 · similarity 0.83)
- Autonomous Policy Explanations for Effective Human-Machine Teaming (AAAI 2024 · similarity 0.82)
- Ad Hoc Teamwork via Offline Goal-Based Decision Transformers (ICML 2025 · similarity 0.81)
- Promoting Research Collaboration with Open Data Driven Team Recommendation in Response to Call for Proposals (AAAI 2024 · similarity 0.81)
- Probabilistic Modeling of Interpersonal Coordination Processes (ICML 2024 · similarity 0.81)
- NestRL: A Nested Training Regime for Mutual Adaptation in Human–AI Teaming (NeurIPS 2026 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.