Binding via Reconstruction Clustering
This addresses the binding problem in representation learning, which is important for improving expressiveness and generalization in machine learning, though it appears incremental as it builds on existing autoencoder and clustering techniques.
The paper tackles the binding problem in disentangled distributed representations where multiple objects in complex data interfere, proposing an unsupervised algorithm using denoising autoencoders with an Expectation-Maximization-like clustering process. The method is demonstrated on artificially generated binary image datasets, showing it can generalize to bind new objects not seen during training.
Disentangled distributed representations of data are desirable for machine learning, since they are more expressive and can generalize from fewer examples. However, for complex data, the distributed representations of multiple objects present in the same input can interfere and lead to ambiguities, which is commonly referred to as the binding problem. We argue for the importance of the binding problem to the field of representation learning, and develop a probabilistic framework that explicitly models inputs as a composition of multiple objects. We propose an unsupervised algorithm that uses denoising autoencoders to dynamically bind features together in multi-object inputs through an Expectation-Maximization-like clustering process. The effectiveness of this method is demonstrated on artificially generated datasets of binary images, showing that it can even generalize to bind together new objects never seen by the autoencoder during training.