A Fast Learning Algorithm for Deep Belief Nets

Citations Over TimeTop 1% of 2006 papers

Abstract

We show how to use "complementary priors" to eliminate the explaining-away effects that make inference difficult in densely connected belief nets that have many hidden layers. Using complementary priors, we derive a fast, greedy algorithm that can learn deep, directed belief networks one layer at a time, provided the top two layers form an undirected associative memory. The fast, greedy algorithm is used to initialize a slower learning procedure that fine-tunes the weights using a contrastive version of the wake-sleep algorithm. After fine-tuning, a network with three hidden layers forms a very good generative model of the joint distribution of handwritten digit images and their labels. This generative model gives better digit classification than the best discriminative learning algorithms. The low-dimensional manifolds on which the digits lie are modeled by long ravines in the free-energy landscape of the top-level associative memory, and it is easy to explore these ravines by using the directed connections to display what the associative memory has in mind.

Related Papers

• → GAN-Leaks: A Taxonomy of Membership Inference Attacks against Generative Models(2020)273 cited
• → Generative Versus Discriminative Learning(2004)26 cited
• → Improved discriminative training for generative model(2009)1 cited
• → Generative Model for Person Re-Identification: A Review(2020)
• → Bridging Generative and Discriminative Noisy-Label Learning via Direction-Agnostic EM Formulation(2023)