Today, in a rainy Las Vegas, we got to find out more on some other new services released at AWS Re:Invent.
One particularly interesting offering is Amazon Kendra, an enterprise search service powered by machine learning. The service indexes all sorts of information such as files, web sites, Sharepoint, databases, etc. and allows the user to search using natural language.
A demo of the system showed a search containing the phrase “Where is the IT Helpdesk”. A normal enterprise search system returns various keyword matches but does not understand that we are looking for the location of the IT Helpdesk. With the same search undertaken on Amazon Kendra it searched some company documents and return some text outlining that the IT Helpdesk was on the 1st floor of the demo company. The machine learning capabilities seem to provide some significant improvements in returning search result which are more relevant to the user. It will be an interesting technology to investigate further.
Another interesting service also introduced and again using machine learning was Amazon Fraud Detector. This service simplifies detecting fraud in transactions. The user provides email addresses, IP addresses, and other historical transaction and account registration data, along with other indicators of fraudulent/non-fraudulent transactions. A machine learning model, trained on this historic data, is then fed new transactions and indicate whether the new transaction is fraudulent or not. The technology is built around the same concepts used by Amazon themselves in their traditional e-commerce business.
Again with Amazon Fraud Detector we’re seeing the packaging of complex machine learning applications into easier to use, packaged services.
Finally another new service which is slightly longer term in it’s ambition is Amazon Braket. This service allows quantum computing algorithms to be programmed and simulated on the cloud. We’re beginning to see quantum computing making progress and services like Braket allow interested parties to get some experience and prepare for this new computing paradigm.