Convolutional Bipartite Attractor Networks
This work addresses the challenge of constructing coherent states from ambiguous evidence in AI, offering a more motivated alternative to deep feedforward models and costly generative methods, though it appears incremental in adapting attractor nets to modern deep learning contexts.
The authors tackled the problem of interpretation in perception and cognition by revisiting attractor networks, proposing a convolutional bipartite architecture with novel training loss, activation function, and connectivity constraints, and demonstrated its potential for image completion and super-resolution on larger problems than previously explored.
In human perception and cognition, a fundamental operation that brains perform is interpretation: constructing coherent neural states from noisy, incomplete, and intrinsically ambiguous evidence. The problem of interpretation is well matched to an early and often overlooked architecture, the attractor network---a recurrent neural net that performs constraint satisfaction, imputation of missing features, and clean up of noisy data via energy minimization dynamics. We revisit attractor nets in light of modern deep learning methods and propose a convolutional bipartite architecture with a novel training loss, activation function, and connectivity constraints. We tackle larger problems than have been previously explored with attractor nets and demonstrate their potential for image completion and super-resolution. We argue that this architecture is better motivated than ever-deeper feedforward models and is a viable alternative to more costly sampling-based generative methods on a range of supervised and unsupervised tasks.