home Artificial Intelligence Stepping up to the AI Testing Challenge

Stepping up to the AI Testing Challenge

The terms Artificial Intelligence (hereafter referred to as AI) or Cognitive Computing are one of the most popular topics in the world of technology. While the concept of AI has been there for a while, (one can imagine that the earliest invention of the computer was an early form of AI), it’s only the recent past that the topic has grabbed more eyeballs than ever before.

Thanks to the increase in computing power of our computers which can today process more than 65,000 MIPS (Million Instructions per Second) and we are just talking about normal computers. This coupled with the ability to analyze huge volumes of big data (thanks to technologies like Hadoop, spark), we now have the ability to more than ever before with the power of computers.

However, the icing on the cake, remains the introduction of the concept called ‘Machine Learning’. Coupled with the ability to analyze huge data sets at great speed, machine learning libraries have enabled the computer programmer to write programs that can ‘learn’ as it executes.

This ability of the computer program to learn as it executes is the most powerful concept of AI. The possibilities of this

“learning” are unlimited. Given the infinite number of possibilities that a computer program can learn and ‘adapt’, this is asking for a whole new perspective of looking at things.

So, what does it mean for the Quality Assurance team? What does it mean to a software testing professional? Well, it certainly is asking for a different way of thinking. While the traditional software test professional had to think about the various (finite) possibilities that can occur as a result of executing a program, today with the power of AI, the possibilities can be infinite (remember the ability to “learn” and “adapt”?)

Sounds scary, right? Don’t be scared. While AI can appear like an overwhelming concept, the fact remains that the ability to learn and adapt by a computer program written using machine learning libraries depends on the ability of the programmer who is writing it.

Moreover, the concept of infinite possibilities cannot be ruled out, it’s not easy to get there by one small machine learning program as it also depends on the type and volume of data being fed to the program for learning.

Having said that, this is a wakeup call for all software test professionals! So, buckle up and wear your thinking hat and get started on this new journey in this new world of cognitive computing!

Facebooktwittergoogle_plusredditpinterestlinkedinmail
Mallarswami Nonvinkere
Mallarswami is a technical professional leading a team of high performing senior consultants, data analysts and senior developers providing technology solutions in the tax domain. He has more than 13 years of industry experience and has been responsible for client, business partner and ISV engagements related IBM Big data solutions alongside enabling them on IBM’s big data technology. He has been a frequent speaker at various international conferences including IBM insights and IDUG Australasia. He has successfully delivered technology boot camps around the world including Asia and America.
Rama Anem
Rama is a Quality professional who managed teams and complex technology-driven projects. She has over 13 years of experience, has worked with companies like IBM, AMD in past, and led global teams. Currently she is responsible for managing multiple software products quality in Sunpower Corporation. She oversaw multiple EDW projects and teams from envisioning through go live. She served as one of the Brand ambassadors for IBM DB2. Rama has contributed in technical software international conferences, workshops and magazines.

One thought on “Stepping up to the AI Testing Challenge

Leave a Reply

Your email address will not be published. Required fields are marked *

Subscribe