A convolutional neural network (CNN) attempts to recognize objects in images but fails to generalize to rotated versions of those objects. Which mechanism explains this failure?

💡 Explanation

The CNN's failure to generalize rotated objects arises because it lacks rotation-invariant feature encoding; it learns features specific to orientations rather than recognizing the object regardless of its orientation. Therefore, the CNN cannot extract the same features when the object is rotated, rather than because of overfitting or regularization issues.

• Why does a literary translator often fail to achieve complete cultural equivalence when rendering a novel?
• Why does a spectrogram, representing spoken language, show smearing of spectral energy when recorded in a reverberant room rather than an anechoic chamber?
• A team is developing a chatbot. If the chatbot incorrectly interprets 'book' as in 'make a reservation' when the user meant 'item to read,' which consequence follows?
• An engineer analyzing a multilingual text corpus notices skewed word frequency distributions. Which mechanism best explains this variation, assuming consistent document sizes?
• If a poet intentionally violates a consistent iambic pentameter with a spondee in an English sonnet, which consequence is most likely?
• Why does a language experience rapid attrition when its speaker base shifts toward digital communication platforms?