MuLab Research

The MuLab focuses on a broad range of research topics, unified by the common goal of better model understanding; both from the perspective of human understading and machine understanding. The models we typically consider are dynamical systems or planning models: i.e., how things can change over time through the interaction of a human/agent with an environment.

A core setting for much of the work surrounds multi-agent reasoning, as many environments require interaction with other agents. Also inherent to many settings is an aspect of uncertainty. In what the agent believes, in what the impact of acting in the world might be, etc. Much of our work embraces the pervasive nature of uncertainty, and aims to address the key challenges it presents.

• Model Acquisition
  • Domain Authoring
  • Principled Induction
  • Theory Learning
• Multi-agent Planning
• Dialogue Agents
• Planning Under Uncertainty
• Execution
• Knowledge Compilation

Model Acquisition

Domain Authoring

• planning.domains
• PDDL book

Principled Induction

• macq
• LTL induction

Theory Learning

• latplan
• latent understanding
• modelling human behaviour

Multi-agent Planning

• ABM

Natural Language & Planning

• PDDL completion
• plan4dial

Planning Under Uncertainty

• PRP & extensions
• RP-MEP
• Autonomous vehicles

Execution

• (Tem)POPEM
• contingent execution

Knowledge Compilation

• macq
• D#
• bauhaus
• python-nnf