Wednesday 28 October 2020

How Can We Trust Testing AI Bots?


The cornerstone of the testing is trust. For the test to be value, we must believe in the output of the test. So how do we believe in testing the AI?

Vighnesh Sreekanta, chief scientist at test.ai, address this question in Automation Guild in 2019 with "Test Machina" talk.

experienced AI

As already mentioned, to use AI effective in testing, we have to get familiar with it. Here are some of the applications and tools that have nothing to do with testing; people can use this to get a sense of what the AI ​​does.

Drawing fast!

Okay, this one's fun.

In a presentation Sreekanta described above, it uses Google's tools. In it, you are challenged to draw objects in 20 seconds. During that time, the AI ​​try to guess what the picture is.

Also Read : Software testing company in Texas

One great way to get familiar with something is to play, and this is a fun tool for it.

taught Machines

Taught Machine is another experiment from Google. This site allows you to try a training model for machine learning. Classifying video samples, photos, or sound to train a machine-learning algorithm. Then give it a try and see how it works.

Of course I had to try this! This site allows using your own video cam to record images. I made a picture holding a solved Rubik's Cube, the cube partially solved, and solved cube, all from different angles. Then I picture are classified into two groups-solved and unsolved.

It was a tremendous job of classifying the new image of the cubes in the new configuration. Of course, as a tester, I can not help but try to confuse it, and I did.

This machine is very well taught to quickly set up the machine to learn and experiment with it. Learn how AI as such can work for you or against you. If you are using AI-enabled testing now or are considering it, this would be a great way to get familiar with the ups and downs of ML.

ML is the most unlimited area for exploratory testing I found.

Know your tools.

Weka

Weka is Waikato Environment for Knowledge Analysis, created and managed by the New Zealand University of Waikato. Use it to explore the data with machine learning. ML You can use different algorithms, configure them, and train them in the dataset.

Weka is the first tool I used to begin to understand ML. While the user interface has room for improvement, do not judge a book by its cover. The strength of this tool goes far beyond appearances.

AI AI in Tests and Testing Meetup - Santa Clara, California

Meetups are a wonderful way to learn, and because of the pandemic, most meetups are now virtual, which means you can join them from anywhere.

The only AI in testing meetup I find is one of the hosts in Santa Clara, California. Last month, they talk about "Creating Deployable Classifier uses Tensorflow Jira Bug." I can not wait to see what's next for this group!

Also Read : Software Testing Company in Bay Area

 

AI tool

The most AI tool is a commercial, for several reasons. Total team effort put into building the AI ​​tool for testing very large, and gifts to customers can be greater than the price many times. Kevin Surace, CTO of Appvance, explains that his company AI tool consists of about 4 million lines of code and took about 250,000 engineering hours to develop.

I believe there will be a lot of AI-based, test-help tool in the open-source market over the next few years. But for now, here are some tools that can help.

 

Test.AI Classifiers

Test.ai classifier using machine learning to match elements on a web page. These classifiers are available in a number of different languages.

 

TensorFlow

Want to use machine-learning API to test your own ideas? TensorFlow is one API to get in there to quickly apply the model ML.

AGENT

AGENT is an abbreviation of "AI Generation and Exploration in the Test." Search this bot generator to test your site on GitHub repo Raja Tariq.

Also Read : Software Testing Company in San Francisco

 

No comments:

Post a Comment