The counterfactual explanations method
In this chapter and section, we will be exploring AI explanations in a unique way. We will not go through the "getting started" developer approach and then explore the code in sequential steps from beginning to end.
We will start from a user's perspective when faced with factual and counterfactual data that require an immediate explanation.
Let's now see which dataset we are exploring and why.
Dataset and motivations
Sentiment analysis is one of the key aspects of AI. Bots analyze our photos on social media to find out who we are. Social media platforms scan our photos daily to find out who we are. Our photos enter the large data structures of Google, Facebook, Instagram, and other data-collecting giants.
Smiling or not on a photo makes a big difference in sentiment analysis. Bots make inferences on this.
For example, if a bot detects 100 photos of people, which contain 0 smiles and 50 frowns,...