We present a very simple method for extending pretrained machine translation metrics to incorporate document-level context. We apply our method to four popular metrics: BERTScore, Prism, COMET, and the reference-free metric COMET-QE. We evaluate our document-level metrics on the MQM annotations…
Spliced binned-Pareto distributions are flexible enough to handle symmetric, asymmetric, and multimodal distributions, offering a more consistent metric.
A combination of cutting-edge hardware, sensor technology, and bespoke machine learning approaches can predict trajectories of vehicles, people, and even animals, as far as 8 seconds into the future.
Subscribe to the Amazon Science newsletter and get the latest news delivered to your inbox each month! The monthly newsletter features in-depth looks at the groundbreaking research and applied science being conducted at Amazon, and introduces you to the scientists and scholars doing that work.
Amazon Science gives you insight into the company’s approach to customer-obsessed scientific innovation. Amazon fundamentally believes that scientific innovation is essential to being the most customer-centric company in the world.