# Low-Shot Learning from Imaginary Data

- the paper "Hallucination Improves Few-Shot Object Detection" is based on ths paper
- this paper uses a meta-learning approach.
- realistic examples might still fail to capture many modes of variation of visual concepts, **while unrealistic hallucinations can still lead to a good decision boundary.**
- training process is **model agnostic.** The hallucination appraoch can be used on different meta-learning methods.
- training using meta-learning allows for learning to hallucinate to make class distinctions (so no need to worry about having to tune for realism or diversity)

## Related Work
- based on the work by Hariharan and GirShick, the effort of annotations can be avoided by **trying to transfer the transformation from a pair of exmaples from a known category to a novel class' seed example**. This paper follows this line of work, in an end-to-end manner.

## Approach

### Meta-Learning

## Experiments
- the benchmark is based on work proposed by Hariharan ad Girshick.
- hyperparameters can be tuned to achieve different trade-off points without substantively changing the average performance.
- one way to embody this trade-off is by incorportating a prior over base and novel-classes.

### Hallucinator

### Evaluation Protocol
1. model is given test samples from the novel classes, and the label space is restricted to Cnovel.
2. model is given test examples from the base clases, and the label space is restricted to base classes.
3. the model is given test examples from both the base and novel classes in equal proportion.

## Discission

Are sophisticated hallucination architectures necessary?
-