On the Transfer of Disentangled Representations in Realistic Settings
This work addresses the problem of evaluating disentangled representations in realistic, correlated settings for machine learning researchers, but it is incremental as it builds on existing methods with new data and architectures.
The paper tackled the scalability and real-world impact of disentangled representations by introducing a new high-resolution dataset with 1M simulated and 1,800 real-world images, and found that disentanglement predicts out-of-distribution task performance.
Learning meaningful representations that disentangle the underlying structure of the data generating process is considered to be of key importance in machine learning. While disentangled representations were found to be useful for diverse tasks such as abstract reasoning and fair classification, their scalability and real-world impact remain questionable. We introduce a new high-resolution dataset with 1M simulated images and over 1,800 annotated real-world images of the same setup. In contrast to previous work, this new dataset exhibits correlations, a complex underlying structure, and allows to evaluate transfer to unseen simulated and real-world settings where the encoder i) remains in distribution or ii) is out of distribution. We propose new architectures in order to scale disentangled representation learning to realistic high-resolution settings and conduct a large-scale empirical study of disentangled representations on this dataset. We observe that disentanglement is a good predictor for out-of-distribution (OOD) task performance.