IPAB Workshop - 04/08/2022

Speaker: Cillian Brewitt

Title: Verifiable Goal Recognition for Autonomous Driving with Occlusions

Abstract: When used in autonomous driving, goal recognition allows the future behaviour of other vehicles to be more accurately predicted. A recent goal recognition method for autonomous vehicles, GRIT, has been shown to be fast, accurate, interpretable and verifiable. In autonomous driving, vehicles can encounter novel scenarios that were unseen during training, and the environment is partially observable due to occlusions. However, GRIT can only operate in fixed frame scenarios, with full observability. We present a novel goal recognition method named Goal Recognition with Interpretable Trees under Occlusion (OGRIT), which solves these shortcomings of GRIT. We demonstrate that OGRIT can generalise between different scenarios and handle missing data due to occlusions, while still being fast, accurate, interpretable and verifiable.

Speaker: Elliot Fosong

Title: Few-Shot Teamwork

Abstract:  We propose the novel few-shot teamwork (FST) problem, where skilled agents trained in a team to complete one task are combined with skilled agents from different tasks, and together must learn to adapt to an unseen but related task.

Solutions to the FST problem could accelerate multi-agent reinforcement learning by reducing the experience required to train a team of agents to complete a complex task such as 5-a-side football by decomposing the complex tasks into simpler sub-tasks (such as defence/offence training drills), training agents to complete these simpler sub-tasks, then forming a team from these agents and using their learned skills to aid learning in the complex task.