NeurIPS 2019 Disentanglement Challenge: Improved Disentanglement through Learned Aggregation of Convolutional Feature Maps

Abstract

This report to our stage 2 submission to the NeurIPS 2019 disentanglement challenge presents a simple image preprocessing method for learning disentangled latent factors. We propose to train a variational autoencoder on regionally aggregated feature maps obtained from networks pretrained on the ImageNet database, utilizing the implicit inductive bias contained in those features for disentanglement. This bias can be further enhanced by explicitly fine-tuning the feature maps on auxiliary tasks useful for the challenge, such as angle, position estimation, or color classification. Our approach achieved the 2nd place in stage 2 of the challenge. Code is available at https://github.com/mseitzer/neurips2019-disentanglement-challenge.

arXiv commentDisentanglement Challenge - 33rd Conference on Neural Information Processing Systems (NeurIPS) - NeurIPS 2019. arXiv admin note: text overlap with arXiv:2002.10003. Acknowledgements added

References in corpus (5)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Auto-Encoding Variational Bayes
• Exact solutions to the nonlinear dynamics of learning in deep linear neural networks
• Disentangling by Factorising
• On the Transfer of Inductive Bias from Simulation to the Real World: a New Disentanglement Dataset