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…
Web crawl scheduling
A web crawler is an essential part of a search engine that visits URLs and downloads their content, which is subsequently processed, indexed, and used by the search engine to decide which results to show…
SALSA
A statistical adaptive procedure to automatically schedule the learning rate for deep learning with a broad family of stochastic gradient methods.
Guidelines for Human-AI Interaction
Developed in a collaboration between Aether, MSR, and Office, these guidelines for human-AI interaction synthesize more than 20 years of thinking and research in human-AI interaction.