IROS 2021 MAIR2 Workshop on

Multi-Agent Interaction and Relational Reasoning

October 1, 2021 | Prague, Czech Republic (Virtual)

14:20 - 16:00 CEST; 8:20 - 10:00 EDT; 5:20 - 7:00 PDT

Welcome to IROS MAIR2 Workshop

Modeling relations and interactions between agents (e.g., robots, humans, entities, objects, and particles) is widely studied and plays an important role in various tasks, necessitating larger-scale communication and collaboration between researchers in different robotic or relevant fields. Our goal is to enable interdisciplinary discussion about multi-agent relational reasoning knowledge from different research areas such as multi-agent reinforcement learning, cooperative and competitive multi-agent systems, intuitive physics, dynamics modeling, object detection and tracking, scene understanding, human-robot interaction. The goal of this workshop is to provide an opportunity to discuss how the concepts and techniques in different fields related to multi-agent interaction modeling could help and advance each other.

Topics Covered

Multi-agent reinforcement learning for multi-robot planning, control, manipulation
Causal discovery and modeling in multi-agent systems
Cooperative and competitive multi-agent systems
Interaction modeling in perception, e.g., object detection, multi-object tracking
Relational reasoning in scene understanding, e.g., scene graph, object segmentation
Relational reasoning in intuitive physics for particles and objects
Multi-robot systems and swarm systems
Human-robot / human-object / robot-robot interaction and collaboration

Evaluation and metrics of interaction modeling
Spatio-temporal relational reasoning in multi-agent systems
Visual transformers and their applications to relational reasoning
Relational representation learning using graph neural networks and its applications
Transfer learning and domain adaptation in relational modeling
Explainability and interpretability in relational reasoning and interaction modeling
Relational reasoning and interaction modeling in trajectory forecasting and behavior modeling of traffic participants (pedestrians and vehicles), sports players, etc.