Text and image explainability with LIME

Explainable AI in Python

Fouad Trad

Machine Learning Engineer

Text-based models

Process and interpret written language
Example: Sentiment analysis
Black box models
LimeTextExplainer explains such models
- Finds how each word impacts prediction

Sentiment analysis model receiving a user review and predicting the corresponding sentiment ('Positive' or 'Negative').

LIME text explainer

from lime.lime_text import LimeTextExplainer


text_instance = 
"This product has great features but a poor design."


def model_predict(instance):
  ...
  return class_probabilities


explainer = LimeTextExplainer()

exp = explainer.explain_instance(

  text_instance,

  model_predict

)


exp.as_pyplot_figure()

Image-based models

Highly complex
Interpret visual data
Example: Food classification
LimeImageExplainer explains such models
- Finds which parts of image impact predictions

Food classification model that receives the image of a pizza and predicts which type of food is in the image.

LIME image explainer

from lime.lime_image import LimeImageExplainer


explainer = LimeImageExplainer()

explanation = explainer.explain_instance(

  image,

  model_predict,

  num_samples=50

)


temp, _ = explanation.get_image_and_mask(

  explanation.top_labels[0],

  hide_rest=True

)

Ice cream image.

LIME image explainer

from lime.lime_image import LimeImageExplainer

explainer = LimeImageExplainer()
explanation = explainer.explain_instance(
  image, 
  model_predict,  
  num_samples=50
)

temp, _ = explanation.get_image_and_mask(
  explanation.top_labels[0], 
  hide_rest=True
)

plt.imshow(temp)

Ice cream image with less relevant parts being obscured with a black color.

Let's practice!

Explainable AI in Python