A2I Home

/A2I Home

The Applied AI Lab (A2I) explores core challenges in AI and Machine Learning to enable robots to robustly and effectively operate in complex, real-world environments. Our research is guided by our vision to create machines which constantly improve through use in their dedicated workspace. In doing so we explore a number of intellectual challenges at the heart of robot learning such as machine introspection in perception and decision making, data efficient learning from demonstration, task-based and transfer learning and the learning of complex tasks via a curriculum of less complex ones. All the while our intellectual curiosity remains grounded in real-world robotics domains such as autonomous driving, logistics, manipulation or space exploration.

Latest News

25/10/2017:

We are hiring for two postdoctoral research positions! If you’re interested in joining our exciting research efforts to push the boundaries in robot learning for perception, action, and interaction, check out the job posts here and here.

15/09/2017:

We are excited to say that we are organising a workshop on “Acting and Interacting in the Real World: Challenges in Robot Learning” at NIPS 2017 in December! This is in collaboration with Raia Hadsell and Martin Riedmiller (Google DeepMind) as well as Rohan Paul (MIT). You can see the details here.

08/09/2017:

A2I will present papers at both CoRL and NIPS!  For further details, see the Research page.

  • M. Wulfmeier, I. Posner, and P. Abbeel, “Mutual Alignment Transfer Learning,” in Conference on Robot Learning, 2017.
  • A. R. Kosiorek, A. Bewley, and I. Posner, “Hierarchical Attentive Recurrent Tracking,” in Neural Information Processing Systems, 2017.

Highlights

2206, 2017

Adversarial Domain Adaptation

Addressing Appearance Change in Outdoor Robotics with Adversarial Domain Adaptation Abstract – Appearance changes due to weather and seasonal conditions represent a strong impediment to the robust implementation of machine learning systems in outdoor robotics. While [...]

704, 2017

Deep Tracking

Deep tracking in the wild: End-to-End Tracking and Semantic Segmentation Using Recurrent Neural Networks Abstract – This paper presents a novel approach for tracking static and dynamic objects for an autonomous vehicle operating in complex urban [...]