Seminar
Towards Lifelong Robot Learning
Abstract: Google Brain Robotics vision is to leverage learning to push the field of robotics forward. As such, we have engaged in research ranging in application from navigation to grasping and approach from deep RL to learning from demonstration. Fundamentally, our research is built around the core idea of lifelong learning. Our long term goal [...]
Exploring Human-Robot Trust during Emergencies
Abstract: This talk presents our experimental results related to human-robot trust involving more than 2000 paid subjects exploring topics such as how and why people trust a robot too much and how broken trust in a robot might be repaired. From our perspective, a person trusts a robot when they rely on and accept the [...]
Carnegie Mellon University
Belief Space Planning for Reducing Terrain Relative Localization Uncertainty in Noisy Elevation Maps
Abstract Accurate global localization is essential for planetary rovers to reach science goals and mitigate mission risk. Planetary robots cannot currently use GPS or infrastructure for navigating, and hence rely on terrain for determining global position. Terrain relative navigation (TRN) compares planetary rover-perspective images and 3D models to existing satellite orbital imagery and digital elevation [...]
Comparing apples and oranges: Off-road pedestrian detection on the NREC agricultural person-detection dataset
Abstract: Person detection from vehicles has made rapid progress recently with the advent of multiple high-quality datasets of urban and highway driving, yet no large-scale benchmark has been available for the same problem in off-road or agricultural environments. In this talk, we present the NREC Agricultural Person-Detection Dataset to spur research in these environments. It [...]
Adaptive Information Gathering via Imitation Learning
Abstract: In the adaptive information gathering problem, a robot is required to select an informative sensing location using the history of measurements acquired thus far. While there is an extensive amount of prior work investigating effective practical approximations using variants of Shannon’s entropy, the efficacy of such policies heavily depends on the geometric distribution of [...]
Deep Structured Models for Human Activity Recognition
Abstract: Visual recognition involves reasoning about structured relations at multiple levels of detail. For example, human behaviour analysis requires a comprehensive labeling covering individual low-level actions to pair-wise interactions through to high-level events. Scene understanding can benefit from considering labels and their inter-relations. In this talk I will present recent work by our group building [...]
Learning Single-view 3D Reconstruction of Objects and Scenes
Abstract: In this talk, I will discuss the task of inferring 3D structure underlying an image, in particular focusing on two questions - a) how we can plausibly obtain supervisory signal for this task, and b) what forms of representation should we pursue. I will first show that we can leverage image-based supervision to learn [...]
From Robust Real-time SLAM to Safe Collision Avoidance
Abstract State estimation plays a critical role in a robotic system. The problem is to know where the robot is and how it is oriented. This is very often a building block in the navigation system, which modules in charge of higher level tasks are relied on. Challenges are to carry out state estimation in [...]
Level Set Models for Computer Graphics
ABSTRACT A level set model is a deformable implicit model that has a regularly-sampled representation. It is defined as an iso-contour, i.e. a level set, of some implicit function f. The contour is deformed by solving a partial differential equation on a sampling of f, an image in 2D and a volume dataset in 3D. [...]