Publications
• Sorted by Date • Classified by Publication Type • Classified by Research Category •
A Behavior-first approach to Multi-Intention Inverse Reinforcement Learning
Antonio Mone, Frans A Oliehoek, and Luciano Cavalcante Siebert. A Behavior-first approach to Multi-Intention Inverse Reinforcement Learning. Transactions on Machine Learning Research, 2026.
Download
Abstract
Inverse Reinforcement Learning (IRL) seeks to infer reward functions from expert demonstrations. When demonstrations come from multiple experts with different intentions, the problem becomes Multi-Intention IRL (MI-IRL). MI-IRL methods generally follow a reward-first paradigm: they ask which reward, if followed, could have generated each trajectory. This leads to trajectory similarity being based on reward likelihood rather than on the relationships between behaviors. In deep generative MI-IRL methods, this paradigm further couples behavior clustering and reward learning, making such methods dependent on prior knowledge of the true number of modes . This limits adaptability to unseen behaviors, and restricts the analysis to learned rewards rather than relationships across behaviors. In contrast, we approach MI-IRL from a different standpoint. We propose Contrastive Multi-Intention IRL (CoMI-IRL), a transformer-based unsupervised behavior-first framework: it first learns a representation of trajectory dynamics that captures intrinsic behavioral similarity, then clusters trajectories in that space, and finally learns a reward for each cluster. Our experiments show that CoMI-IRL achieves higher clustering and reward learning performance when compared to Ess-InfoGAIL without needing or labels, while allowing for visual interpretation of behavior relationships and adaptation to unseen behavior without full retraining.
BibTeX Entry
@article{Mone26TMLR_behavior-first,
title={A Behavior-first approach to Multi-Intention Inverse Reinforcement Learning},
author={Antonio Mone and Frans A Oliehoek and Luciano Cavalcante Siebert},
journal={Transactions on Machine Learning Research},
issn={2835-8856},
year={2026},
url={https://openreview.net/forum?id=42kUyc9KGs},
note={},
keywords = {refereed},
abstract = {
Inverse Reinforcement Learning (IRL) seeks to infer reward functions from
expert demonstrations. When demonstrations come from multiple experts
with different intentions, the problem becomes Multi-Intention IRL
(MI-IRL). MI-IRL methods generally follow a reward-first paradigm:
they ask which reward, if followed, could have generated each
trajectory. This leads to trajectory similarity being based on reward
likelihood rather than on the relationships between behaviors. In deep
generative MI-IRL methods, this paradigm further couples behavior
clustering and reward learning, making such methods dependent on prior
knowledge of the true number of modes . This limits adaptability to
unseen behaviors, and restricts the analysis to learned rewards rather
than relationships across behaviors. In contrast, we approach MI-IRL
from a different standpoint. We propose Contrastive Multi-Intention
IRL (CoMI-IRL), a transformer-based unsupervised behavior-first
framework: it first learns a representation of trajectory dynamics
that captures intrinsic behavioral similarity, then clusters
trajectories in that space, and finally learns a reward for each
cluster. Our experiments show that CoMI-IRL achieves higher
clustering and reward learning performance when compared to
Ess-InfoGAIL without needing or labels, while allowing for visual
interpretation of behavior relationships and adaptation to unseen
behavior without full retraining.
}
}
Generated by
bib2html.pl
(written by Patrick Riley) on
Thu Oct 08, 2026 13:51:42 UTC