Geometric Disentanglement by Random Convex Polytopes
This work addresses representation evaluation for machine learning practitioners, but it is incremental as it builds on existing methods like k-means clustering.
The authors tackled the problem of measuring representation quality in deep learning by proposing the Random Polytope Descriptor, a geometric method using random convex polytopes, and demonstrated that regularization in autoencoders can decrease out-of-distribution detection performance.
We propose a new geometric method for measuring the quality of representations obtained from deep learning. Our approach, called Random Polytope Descriptor, provides an efficient description of data points based on the construction of random convex polytopes. We demonstrate the use of our technique by qualitatively comparing the behavior of classic and regularized autoencoders. This reveals that applying regularization to autoencoder networks may decrease the out-of-distribution detection performance in latent space. While our technique is similar in spirit to $k$-means clustering, we achieve significantly better false positive/negative balance in clustering tasks on autoencoded datasets.