IBM’s toolbox helps your AI survive in real world

Creating an artificial intelligence is definitely not a business-as-usual. It's an art in itself. Most advanced AIs utilise deep neural networks (DNN) - machine learning models - to achieve human-level performance on cognitive task such as image recognition. However, DNNs are not invincible. They are sensitive to certain types of attacks, such as adversarial examples. In short, AE attack happens when DNN receives an input, that was tweaked to modify (otherwise accurate) response, thus misclassifying object on a picture.

According to IBM Security Systems CTO Sridhar Muppidi:

One of the biggest challenges with some of the existing models to defend against adversarial AI is they are very platform specific. The IBM team designed their Adversarial Robustness Toolbox to be platform agnostic. Whether you’re coding/developing in Keras or TensorFlow, you can apply the same library to build defenses in.

source: thenextweb.com

So what one can do to prevent this?

IBM Research Ireland released the Adversarial Robustness Toolbox: a library dedicated to adversarial machine learning. Its purpose is to allow rapid crafting and analysis of attacks and defence methods for machine learning models. The Adversarial Robustness Toolbox provides an implementation for many state-of-the-art methods for attacking and defending classifiers.

Read more @ https://www.ibm.com/blogs/research/2018/04/ai-adversarial-robustness-toolbox/

image source: https://techcrunch.com/