One of the largest features that our platform uses is a system of Reviews Automatic Filtering (RAF) with a key component, which is IBM Watson. As you know, the foundation of Revain is smart-filtering reviews so that we can successfully fight off bots, malicious, and plain frivolous users. Let’s take a more detailed look at how that is done.
There are two stages of review filtering. The first is passing the checks by the RAF, which will have something to do with basic things like keyword analysis and more complicated things like pattern recognition by IBM Watson and the RAF on the whole.
One of the largest features that our platform uses is IBM Watson (as part of the RAF), which will be used to smart-filter the reviews which our company is based on. Today we are going to take a look at the strengths our IBM Watson can offer in safeguarding your reviews so your reviews can safeguard the crypto community.
We do hope that a lion’s share of our users at least have read and understood the terms and conditions Revain offers, and we hold some hope that all our users know we have implemented an AI that offers additional protection.
While most people have a vague idea of what it’s like to play chess with a computer at level 9, few actually know what an AI is and how it can help. So let’s go through things methodically, step by step.
What is it?
IBM Watson, which has been created by, you guessed it, IBM, is an artificial intelligence program with a focus on deep learning that was created in 2014.
IBM Watson seems to be permeating every area of our lives now, slowly but surely at the beginning and then picking up speed exponentially.
Versions of it that are used today and are tailored to specific companies and niches are far more complex, however, than most imagine, consisting of a Tone Analyzer, market business analysis, voice-activated commands, and more. Revain’s AI consists of IBM Watson, but also of parts specifically designed and tailored to the system by developers.
Features like keyword recognition and natural language understanding are an addition to the system which will allow it to fight of bots and scammers using patterns analysis. Natural language understanding and the ability self-educated are features scheduled for release, but they are being worked on right now.
How does it work?
While we encourage you to deep-learn the whitepaper (plus what makes our task more difficult is that some of the details are hidden for security reasons), the essentials are as follows:
Automatic review filtering, which is used by some major tech companies like Google and Facebook, is based on elements such as:
Tone analysis
For example, if a review is constructive and the levels of Anger, Sadness, and so on are too high, the review gets turned down.
Natural Language Understanding (in the making)
Analyzing semantic blocks, keywords, and so on, in order to detect whether the text was submitted by a human or a machine based on learned patterns of human input (below).
Learning to learn
In the same way that DeepMind has learned from scratch to outdo even the most advanced gamers, IBM Watson compares its data and patterns with the reviews submitted by users, so it can perfect itself and evolve. For security reasons we are unable to divulge exactly how it works, but the RAF is able to analyze user input, recognize patterns, learn, and react to following user input, thus making itself more advanced.
We hope you are duly impressed by our work and, as always, we eagerly look forward to all your feedback and participation.