Cogniac’s Pipeline

Original Source Here

Cogniac’s Pipeline

Amy Wang, Cogniac Co-Founder and VP of Systems, shares how Cogniac’s proprietary pipeline is being used as an enterprise solution.

One of the things I’m really proud of at Cogniac is the breadth of industries we serve. Through our Visual Intelligence Platform, we have enabled companies to harness the power of AI without having to bring on their own data scientists or computer vision experts. One of the features of the Cogniac platform that makes this possible is our proprietary Cogniac pipeline — a workflow of individual tasks that users can build up to analyze their visual data.

We refer to Cogniac as “Our Superhuman Vision” because we help people see things faster and more accurately than even human subject matter experts. Alongside our customers, we’re taking on challenges that they (and we) never thought possible, and this is being enabled by our no-code AI platform. It is this pipeline that underpins the entire system, really, and it allows users all around the world to create the workflows that matter to them.

When a customer wants to analyze images for a task, for example, reading the numbers on hazmat placards on a train, the platform helps them to break down the steps required to get to that data. Sometimes it can be just one action but in other cases, it might be three or four. The reason we break these steps down is so that the user can select ‘applications’, as we call them, that make sense for them. It might be Optical Character Recognition (OCR), object detection, or classification — whatever is required for the customer to produce the analysis desired. At this point, the user’s technical ability tends to expire and Cogniac can take over, so in the case of the hazmat placards, we create an application to detect the placards then an OCR application to read the numbers displayed. It’s a really simple process that’s actually based on how humans process information.

Underpinning these individual ‘applications’ is a combination of AI techniques that deliver the work required. We deploy hyper-parameter optimization (HPO) across our convolutional neural networks so that when subject matter experts validate our system’s analysis of an image it learns and improves. We call this process ‘AI that creates AI’. This is the snowball effect of our AI — once you kick it off, it keeps learning, improving, getting faster, and more powerful. This means that real-world challenges are now being met by enterprise AI, rather than just academic AI theory.

I always wanted to develop my skills and work with the best minds in the industry to create something truly ground-breaking, something that could change the world. In Cogniac and its pipeline, we are driving change in so many areas that will have a real impact on the way we work — AI is incredibly powerful, and we are proud to be placing it in the hands of companies who are constantly striving to be better.


Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: