Why adding web photos can improve a plant-identification app
Case context: A team building a mobile app to identify flower species has 8,000 photos taken by users on phones. They add 25,000 flower images collected from gardening websites to the training set.
Question: Using the idea of shared patterns between data sources, explain why the extra website images are likely to improve performance on the phone photos.
Sample answer: The website images provide many additional examples of flowers, so the model sees more variation in petals, colors, and shapes. Because the website photos and the phone photos still show the same kinds of flowers, the model can transfer what it learns from the larger website collection to the smaller phone-photo set. That extra coverage usually helps the classifier make better predictions on user photos.
Key points:
- The larger dataset gives the model more flower examples.
- The two sources are similar enough for knowledge to transfer.
- The added data should improve classification on phone photos.
Rubric: Full credit for explaining both that the model gets more examples of flowers and that the website and phone photo distributions overlap enough for the learned features to help on user photos.
0
1
Tags
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Data Science
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
Why can a model trained on street-level dog photos still help on pet-app photos?
True or False: Adding more web images can give a classifier more examples of what a bicycle can and cannot look like.
Online store photos and customer-submitted photos share some _____.
Match each component to its role in transferring knowledge from a broad source set to a target app.
Arrange the reasoning behind why extra catalog photos can improve a product classifier for a shopping app.
Why can catalog photos help a model trained for shopper photos?
Why adding web photos can improve a plant-identification app
Why can examples from one photo source help train on another?
What is a likely result of adding 20,000 extra online photos to improve a traffic-sign classifier?
True or False: A neural network trained on photos from a document scanner can still transfer useful features to similar photos taken with a phone camera.