Explainable AI is a commonly heard term since the last few years. The incoming of this ExAI changed the way ML models are interpreted and understood by ML Practitioners and industries. This concept has extended the definition of how ML model works — From Input, Model and Output to why we receive a specific value and output and which input contributed to it. In this article, we will discuss Explainable AI, what led to the creation of AI and it’s application side.
Tracing the history of AI
In traditional AI, there is no concrete explanation for linear models on why we receive a specific output. Initially, AI models were considered as ‘black-box’ models where there is no clear definition on it’s reason for getting a specific output except the math that works under the hood. The whole ML System was depicted as a three component diagram — Input Features, Model and Output.
What is Explainable AI
Let us now look at how Explainable AI makes ML models more transparent and understandable. Explainable AI helps to understand which input contributed more to the output. This way we can get an insight into which feature is causing the specific output value and which features are having less contributions. This becomes really necessary for models which solve critical problems and which involves many feature values. This technique makes ML models more ‘white-box’ and interpretable.
Application side of ExAI
Let us have a look into the application side of Explainable AI. Consider a heart disease prediction system. A large contribution of single medical parameter out of many medical values might trigger as a heart disease prediction but it can be a large variation that is not necessarily a heart disease. Such cases call for white-box techniques that helps medical experts to understand what caused the model to trigger it as a heart disease.
Thank you for reading!
Useful Links:
Find me on LinkedIn: https://www.linkedin.com/in/vishnuu0399
Know more about Me: https://bit.ly/vishnu-u
