Amazon Silk
Developer Guide

Machine Learning

All browsers make an effort to optimize performance over time. A simple example is device-side caching, wherein the browser temporarily stores requested resources so that subsequent requests can be fulfilled more efficiently. However, many optimizations are limited because they’re based only on individual browsing behavior and because they tend to be ephemeral. For example, clearing browsing data cancels the performance gains made by caching.

By combining machine learning with the resources available in the Amazon cloud, Amazon Silk realizes new possibilities in optimization. Amazon Silk can preprocess and analyze web sites and make dynamic decisions about how to deliver pages to provide the best possible browsing experience.

Because so much of the work of delivering web pages happens on the Amazon cloud, Amazon Silk's optimizations are informed by massive amounts of anonymous, aggregated web traffic data. In effect, Amazon Silk can learn from web traffic patterns and use this learning to improve the individual user experience.