Weighted Meta-Learning
Microsoft Search, Assistant and Intelligence
Leading Microsoft’s effort to build deep intelligence into the products that millions of people use everyday.
VQA Introspect
This projects rethinks the way how VQA models need to trained by proposing to link reasoning questions with simpler, sub questions that are required to learn and solve complex tasks. The work introduces a new…
Inside look at the DARPA Subterranean Urban Circuit Challenge 2020
Over the last year, ten teams from around the world have been participating in the DARPA Subterranean challenges. These challenges are designed to test drones and robots on how they perform in hazardous physical environments…
Learning Visuomotor Policies for Aerial Navigation Using Cross-Modal Representations
Machines are a long way from robustly solving open-world perception-control tasks, such as first-person view (FPV) aerial navigation. While recent advances in end-to-end Machine Learning, especially Imitation and Reinforcement Learning appear promising, they are constrained…