GAB AI is our experimental Beta natural language processing engine. It is a truly big data-based, highly scalable system. It is an entirely new way of computing language. We extract the meaning behind the sentence behind the text without any ‘keyword’ type of search.
The central premise of GAB AI is what we call, ‘Topic of Interest’ and we are testing how these kaleidoscope of ‘topics of interest’ can help understand the meaning behind text.
Actionable Intelligence
GAB AI
One of the most important aspect of language is the context as the written language is highly subjective and contextual. Determining context can give us unprecedented insight into meaning or Topic of the text.
Scraping all social platforms and pulling out a snapshot of all the chatter can be done literally However just listening is no longer enough, we need to understand social media in a context.
0 1
Every subject is written in relation with others, in other words no subject is written in an isolation.
0 2
There is always a ‘Main Topic’ and other Reference or Sub-Topics, which we detect. This allows us not only to analyze ‘Context’ or ‘Topic’ but go a step closer to it’s meaning.
0 3
We go through every sentence of every discussion/article/conversation and predict out of 1000s relevant topics, which one is mentioned. This gives us comprehensive view of not only main topic but also referenced topics.
0 4
For example, We will be able to tell not only that there are more discussions on the topic of ‘Unemployment’ but also that largely it is in reference with ‘Reservation Quota’.
How does it work ?
Meaning out of Text
Our A.I. techniques look at a sentence beyond just a bag of words and look at the context/semantics of that sentence. Not only that but every time new relevant text is consumed our A.I. becomes better and better at its task.
How does it work ?
Meaning out of Text
As shown above, we can see that Unemployment is being discussed mainly from the point of view of ‘Reservation Quota’. We will have insights on different discussion points for each one of the 1000s of topic.