September 1, 2017

Vuukle can now identify Comment Spam while its being typed!

The Honeypot Project received 3,506,236 unique spam messages in just this week. The spamming industry makes gross worldwide revenues to the tune of $200 million per year. Even the conservative estimates state that spamming costs American firms and consumers almost $20 billion.Vuukle has since its inception taken the fight against Spam very seriously and has armed itself with multiple anti-spam features. Now, it can also identify comment Spam even as it is being typed.

The ratio of external costs to internal benefits of fighting spam on your own is around 100:1. This is a very high externality ratio. This skewed up ratio makes it pretty costly for individual publishers to take down spam on their own. It is important to note that publishers can individually either give quality focus on content or on tech development to tackle Spam.

Vuukle being a tech-focused company with the intent to solve publisher pain-points has taken up the fight against Spam vigorously.

The Vuukle Anti-Spam Arsenal:

Vuukle has slowly worked on fighting this menace and has managed to upgrade its anti-spam arsenal to a great extent. Right from suspicious IP address tracking, keyword blocking to an auto-pilot moderation panel, all of these work together. The latest addition is an Artificial Intelligence API. Backed by machine learning API Vuukle can now identify Spam even before the comment is posted. Click To Tweet

Vuukle Comments can identify comment Spam even as it gets typed.

The normally used methods of preventing spam are to block certain keywords, have them fill up a Name and email field, etc. These however are not enough to cease the ever-evolving tactics of Spammers.

With so much money involved spammers always adapt to the changes and somehow evolve their play. This makes it necessary that the tool and technology used against spam develops and evolve by itself as well.
This is exactly where a machine learning supported Artificial Intelligence Spam filter comes into play.

Understanding Machine Learning:

Machine learning is something that gives systems an ability to learn without being explicitly programmed. It relies on deep learning neural networks which allow for integrated growth in its tactics and evolving by learning from previous mistakes and massive amounts of available data.

Thus, such systems when deployed on the web can generate behavioral profiles, spot trends in big data analysis, spot meaningful anomalies and make predictive analysis.

The AI adapts and evolves by itself. It will get better to identify Comment spam.

This AI system can thus adapt and evolve as it grows and more data comes its way. Its pretty good at identifying spam links in comments right now and sends them for moderation. It maybe will be able to specifically segregate spam and take the decision on aggressive spam on its own as it learns.

The Difference this AI makes:

Right now most of the bulk spam is bot generated. The bots autofill the Name and Email category and post a spam comment. In rare cases it is by intended human effort of typing and posting a comment.

What happens after deploying this tool is that, it scans the comment for spammy even while being typed. When one hits the Post button, the comment verification once again takes place. If the comment contains Spam, the user gets a notification stating there seems to be spammy content in the comment.

It sends a notification to the reader to either modify the spammy content or choose to let it through. It still will have to pass the moderation check later.

Now here if the user is human, they would most probably modify the comment or else hit Post if they consider it to be OK. But in the case of bots the scenario changes.

Bots hit rapid comments on target sites and move on. They wouldn’t respond to the Spam notification and thus the comment would never get posted. Thus, this AI in Vuukle Comments tackles bot spam pretty efficiently. Click To Tweet

Moderation becomes Super Easy:

Thus, indirectly the system manages to eliminate most of the bot spam. All of this happens while the comment hasn’t even been submitted. Now suppose by human action or by an advanced bot system the Spam comment gets posted even after notifying, it still has to pass through the moderation panel.

Analyzing the comment for Spam within seconds before allowing it to get posted.

The publisher can at anytime either remove and allow comments that have been brought up for publishing. The AI system will learn from this behavior and it will be able to segregate spam even better.

What this does is reduces enormous moderation load at the outset itself.

Instead of moderating humongous amounts of spam comment, a huge number of these get eliminated even without human intervention from the publisher.

This is in addition to all the tools that Vuukle anti-spam arsenal already possesses.


Vuukle has and will always take the fight against Spam seriously. We will bring out whatever tools are needed to get this right. This latest anti-spam AI addition is a pretty powerful feature. A feature which will only keep getting better with time, learning & evolving on its own.

Santoshkumar Pandey

Engineer by qualification, content & marketing expert by profession. Meditate, blibliophile & a creative geek.