Age guessing with Microsoft is FUN! Now give us your metadata – The Register

Microsoft has unwrapped a metadata-slurping website powered by user-uploaded pictures which pretends to be a fun age-guessing game.

While the page’s inability to accurately guess users’ ages might seem an amusing cock-up on Redmond’s part, the machine learning boffins who set up the page are content to slurp up the JavaScript Object Notation metadata of the photographs users are uploading, and process it to advertise Microsoft’s services.

An invitation extended to the internet-at-large to help Satya Nadella sell his services to all comers has been duly answered by social media users, ever-keen to jump on the new meme bandwagon as early as possible.

In a blog post authored by Corom Thompson and Santosh Balasubramanian, engineers in information management and machine learning at Microsoft, the pair claim they launched a webpage at how-old.net to “play” with a newly released face detection API. The page allows users to upload a picture and have the API attempt to predict the age and gender of any faces recognised within it.

They explain that traffic began to pick-up after “we sent email to a group of several hundred people, asking them to try the page for a few minutes and give us feedback – optimistically hoping that at least 50 people would give it a shot. We monitored our real-time analytics dashboard to track usage and, within a few minutes, the number of people using the site vastly exceeded the number of people we had sent our email to”.

Within a few hours, the post notes, over 35,000 users had visited the page from all over the world in “a great example of people having fun thanks to the power of [Machine Learning]!”

Fun indeed, as Twitter users seemed to visit the page as often for the sake of a joke as they did to receive some valuation of their appearance, with responses to inaccurate age estimations strongly correlating with users’ capacity for humility.

Thompson and Balasubramanian explain “the magic behind this” in a post revealing the page to be a laudably clever advertisement for Microsoft’s services. “It took a couple of developers just a day to put this whole solution together, starting with the pipeline from the web page to the Machine Learning APIs to the real-time streaming analytics and real-time BI”.

Going in depth to advertise those services even more, the writers explain how “in addition to age and gender, we also got […] information such the User Agent string of the users of the web site, the latitude and longitude of location from where the picture was uploaded and more”.

They example this in a representative JSON document, which notably mentions a competitor’s product.

[ { “event_datetime”: “2015-04-27T01:48:41.5852923Z”, “user_id”: “91539922310b4f468c3f76de08b15416″, “session_id”: “fbb8b522-6a2b-457b-bc86-62e286045452″, “submission_method”: “Search”, “face”: { “age”: 23.0, “gender”: “Female” }, “location_city”: { “latitude”: 47.6, “longitude”: -122.3 }, “is_mobile_device”: true, “browser_type”: “Safari”, “platform”: “iOS”, “mobile_device_model”: “IPhone” } ]

Once the pair extracted the information they wanted from the uploaded pictures and web logs, they then collected and analysed the data obtained from the thousands of users uploading pictures to the site in real time. They advise readers to note that, although the information is collected when the user uploads a picture, the picture itself is not saved, it is merely the metadata extracted in the JSON file that gets streamed to their service hub. The information was then aggregated and processed, for no apparent purpose other than advertising a stream processing service.

The Register has contacted the blog’s authors to find out whether the page was sponsored by Microsoft or if it was a personal project. We also asked whether users’ data will continue to be collected and processed, and for what purpose – beyond clever advertising. ®

Sponsored:
Benefits from the lessons learned in HPC

Comments

Write a Reply or Comment:

Your email address will not be published.*