Multi-Agent Goal Recognition with Team- and Goal-Conditioned Reinforcement Learning and Factorized Branch-and-Bound

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Thiago Thomas, Gabriel de Oliveira Ramos, Felipe Meneguzzi
• Venue: arXiv: cs.MA
• Published: 2026-06-24
• Collected: 2026-06-25
• URL: http://arxiv.org/abs/2606.25978v1
• PDF: https://arxiv.org/pdf/2606.25978v1
• DOI: UNKNOWN

Plain-Language Brief

The paper presents Multi-Agent Goal Recognition with Team- and Goal-Conditioned Reinforcement Learning and Factorized Branch-and-Bound (MAGR-BB), a method to infer which agents in a multi-agent system are working together and what their goals are, based on the agents' trajectory. This method is designed for real applications like drone surveillance and collaborative robotics where the agents' actions can be observed but not their specific intentions. MAGR-BB outperforms exhaustive search methods in a controlled multi-agent Blocksworld benchmark by significantly reducing the time required to recognize team goals.

What the Document Says

• Multi-Agent Goal Recognition with Team- and Goal-Conditioned Reinforcement Learning and Factorized Branch-and-Bound (MAGR-BB) is presented.
• This method is used to infer team goals of agents based on their trajectory.
• It is designed for real applications like drone surveillance and collaborative robotics.
• MAGR-BB outperforms exhaustive search methods in a controlled multi-agent Blocksworld benchmark.
• It reduces the time required to recognize team goals substantially.

Key Concepts Mentioned

• Multi-Agent Goal Recognition
• Branch-and-Bound
• Team- and Goal-Conditioned Reinforcement Learning
• Factorized Branch-and-Bound
• Controlled multi-agent Blocksworld benchmark

Claims Made by the Document

• MAGR-BB outperforms exhaustive search methods.
• MAGR-BB reduces the time required to recognize team goals substantially.

Evidence Used

• On a controlled multi-agent Blocksworld benchmark, MAGR-BB returns the same top-ranked hypothesis as exhaustive search throughout the trajectory while cutting hypothesis materialization by orders of magnitude and reducing cumulative recognition runtime substantially.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Multi-agent goal recognition asks an observer to jointly infer which agents act together and what each team is trying to achieve, so the hypothesis space grows combinatorially with the number of team partitions and goals per team. Real applications such as drone surveillance and collaborative robotics expose only the agents' trajectory, which forces the observer to rank team-goal hypotheses from behavior alone. Multi-Agent Goal Recognition with Branch-and-Bound (MAGR-BB) addresses this setting with a shared team- and goal-conditioned policy used as the scoring model inside a factorized branch-and-bound search. On a controlled multi-agent Blocksworld benchmark, MAGR-BB returns the same top-ranked hypothesis as exhaustive search throughout the trajectory while cutting hypothesis materialization by orders of magnitude and reducing cumulative recognition runtime substantially.