Amazon-UCLA model wins coreference resolution challenge
Amazon just beat everyone at teaching AI to understand "that thing over there."

Why it matters
Amazon's breakthrough in coreference resolution could give them a competitive edge in multimodal AI interfaces, making voice assistants and smart devices more intuitive for consumers.
The key facts
4 to knowAmazon-UCLA collaboration won coreference resolution challenge
Model maps spoken language to objects in images
Focus on multimodal device communication
Competitive benchmark victory against other AI models
Go to the source
Amazon Scienceamazon.science
Publisher excerpt: Models that map spoken language to objects in an image would make it easier for customers to communicate with multimodal devices.