Multi Task Learning with Multi Class Classification

Curious to know if you could also have a multi task learning with multi class classification. Are there any examples to understand this?

What does “multi task learning” mean?

Hey @devisdeepai,

Well i guess yes there is. For example, consider a scenario where you have a dataset containing images of various fruits, and you want to build a model that can classify the fruits into their respective categories (e.g., apple, banana, orange) and also predict the ripeness of each fruit (e.g., ripe, unripe, overripe). This is a multi-task learning problem where the model is trained to perform two classification tasks: fruit type classification and ripeness classification.

1 Like

Thank you, I was more looking for a real life example in something like an autonomous car

Very high level, it refers to having one network to learn several things at the same time, where each of the tasks learn and benefit from each other. The example given is the autonomous vehicle, where the goal is to learn to identify what are all the objects around that are present: egs, pedestrian, stop sign, other cars and traffic lights. Please refer to the videos and lessons for detailed information