CleanSpeak 3.18 Update Even More Powerful
- By Bryan Giese
- CleanSpeak, News
- June 14, 2018
It gives us #@$%!* pleasure to announce the CleanSpeak 3.18 update is available! This update takes advantage of full support for regular expressions in Blacklist phrases, as well as enhancements to our API validation. CleanSpeak 3.18 also includes a few minor housekeeping updates that weren’t causing any issues, but it made us feel more comfortable knowing that they would never be able to. Clean code is happy code.
Going Beyond Simple Profanity
Out of the box, the CleanSpeak Blacklist filter provides an industry-leading profanity discovery and filtering process that catches the majority of issues that need to be reviewed or fully blocked. We handle leet speak, phonetic replacements, repeat characters, complex embeddings, and much more.
Unfortunately, there are always people who pride themselves on finding new ways to circumvent filter restrictions. They craft unique situations and creative trolling strategies that we need to address with a specific filter. One evolving challenge has been the spam-marketing style phrases that tend to get posted in very active chats, forums and message boards—i.e. “Buy Fac3book likes” or “By 1ooo Tvvitter F0ll0wers”. These are missed by standard filters because the individual words and terms aren’t considered offensive even though the intent is a violation of the terms of service. It is difficult to understand the context and remove these types of phrases from the stream.
Catching Obfuscated Phrases with CleanSpeak 3.18
A rudimentary solution would be to scan content for these terms just like any other offensive word, although with all the possible variations, that would be extremely difficult and inefficient. Here are a few examples that mean the same, but all possible variations would be difficult to scan for:
- Buy 1,000 followers
- By looo f0llowers
- Bi one thousand pholl0wers
- Byy 1 th0us4nd ph0llowers
With the advances in the CleanSpeak 3.18 update, we are able to catch complex phrases like these as well as detect future patterns as they evolve. How do we do it? I can’t share all the details but it’s a combination of several strategies. We start with our already powerful Blacklist filter and add advanced pattern matching using regular expressions. Then we include a layer of sophisticated concept tagging that focuses on the meaning of the terms. The combination of these components (plus a dose of Inversoft magic) creates a highly scalable solution that amplifies the power of the individual elements. This update increases the accuracy and flexibility of CleanSpeak’s Blacklist filter and gives it advanced capability to identify complex and deliberately obfuscated phrases that are becoming more common. We know not every client will need this level of control in their filtering, but for those that do, it is a welcome new addition to CleanSpeak’s toolbox.
For the full list of updates and most recent information about CleanSpeak, sign up for our newsletter below or visit the CleanSpeak Release Notes here. Please contact us if you would like to learn more about the most effective profanity filtering strategies on the market.
Learn More About CleanSpeak
CleanSpeak is a complete profanity filtering and moderation solution that assists businesses in quickly and easily managing their online communities and any user-generated content. Find out more about CleanSpeak and sign up for a free trial today.