Developer Tech Minutes: Human Parity on CommonsenseQA

From the creator

Microsoft’s KEAR model (Knowledgeable External Attention for commonsense Reasoning model) has achieved human accuracy on the Commonsense Question Answering leaderboard. Discover how researchers from the Azure Cognitive Services team leveraged knowledge from varied external sources to inject commonsense knowledge in deep learning models and, thereby, achieve this incredible milestone. Presented by Yichong Xu, Senior Researcher, from the Azure Cognitive Services Research team. To learn more about the KEAR model, read the latest research publication (https://www.microsoft.com/en-us/research/publication/human-parity-on-commonsenseqa-augmenting-self-attention-with-external-attention/) or the blog post announcing its achievement (https://aka.ms/kear-blog). You can also visit our Tech Minutes webpage for additional resources: https://innovation.microsoft.com/en-us/tech-minutes-human-parity-commonsenseqa Finally, you can discover all the latest Tech Minutes for Developers through our YouTube playlist (https://youtube.com/playlist?list=PLlrxD0HtieHjNJcVCIktce-wSFZ5svwA3) or the Innovation Tech Hub (https://innovation.microsoft.com/developer)

Published 2021-12-20 · Watch on YouTube ↗
Availability and playback restrictions are controlled by the original channel.