1: \begin{abstract}

2: This paper presents our experiments to quantify the manifolds learned by ML models (in our experiment, we use a GAN model) as they train.

3: %We compare these epoch manifolds to the real manifolds representing the real data.

4: We compare the manifolds learned at each epoch to the real manifolds representing the real data.

5: 

6: %To quantify a manifold, studied what metrics we extract from the manifold learned by the ML model, how these metrics change as we continue to train the model, and how we can measure the convergence of the model using manifold metrics.

7: 

8: To quantify a manifold, we study the intrinsic dimensions and topological features of the manifold learned by the ML model, how these metrics change as we continue to train the model, and whether these metrics convergence over the course of training to the metrics of the real data manifold.

9: 

10: %We believe the same set of metrics can be used to compare two models - i.e. are two ML models equivalent? overlap? etc. Furthermore, these metrics can inform the training process as a new cost function.

11: \end{abstract}

12: