Recently we have been testing a new component to our spam filtering system. This component is powered by Cloudmark, one of the most popular and powerful spam filter systems available. We would now like to make Cloudmark available to more customers.
How Cloudmark works
Cloudmark is designed to detect known spam better and works as a central authority based on reporting by millions of Cloudmark users. It would help us improve our implementation of Cloudmark to have more Runbox users testing it.
Customers who are testing Cloudmark don’t need to do anything different in the way they use their email. However, we ask testers to report spam (or genuine mail) that is not classified correctly to a special Runbox email address.
No data is shared with a third party when using Cloudmark, as it’s running on Runbox’ own servers. Any reporting done by our customers is currently only going to Runbox itself. When we implement a reporting facility back to Cloudmark in the future it will be implemented as a clearly marked option.
How to start using Cloudmark
If you are interested in having Cloudmark added to your account, or wish to ask questions about it, please let us know at Runbox Support (support@nullrunbox.com).
This sounds promising, and I’m glad to see that Runbox is leaning on a well-established system rather than trying to go-it-alone. There’s advantages to both approaches, but in the case of spam filtering, accuracy is paramount, so utilizing Cloudmark is probably the better route.
Regarding:
“However, we ask testers to report spam (or genuine mail) that is not classified correctly to a special Runbox email address.”
I volunteered to test. What is the special Runbox email address to report false negatives/positives?
Replied to your support ticket about this 🙂
Six weeks into training Cloudmark… how’s it going?
So it seems you’re not longer taking voluntary spam reports. I got the auto-reply, yet it failed to provide sufficient details on: (1) when the new training system will be in place, and (2) how you will deal with spam that’s not being caught in the meantime, or false-negatives. Should we be using Webmail to assist with the training? Or just wait for your next step? Please clarify.
The auto-reply was just to let you know that we no longer required the spam reports, nothing more. We felt it fair to do that rather than people waste their time. We are very grateful for the reports so far as they have been useful.
The next phase is launching training via the web interface, and then later we will implement training via IMAP. The latter is something we have never had before even for the existing trainable spam filter, and it will be a welcome feature for those who rarely or never use the webmail.
We will make announcements as appropriate when these features are available.
Im getting lots of spam emails now, many times than previous so its much worse for me now.