Adapting a Model for a Rare Class

A large model, pre-trained on a massive, diverse dataset, is tasked with a new classification problem: identifying a specific, newly discovered species of plant. The model is only given a single labeled image of this new plant. Which statement best analyzes the underlying principle that allows the model to potentially succeed at identifying other images of this plant?

Feasibility of One-Shot Learning for a Niche Task

A model's success in a one-shot learning scenario (e.g., classifying a new type of animal after seeing only one image) is fundamentally enabled by the model having previously learned a rich, well-organized feature space where new categories can be easily distinguished, even if it has never seen that specific category before.