Deep learning via message passing algorithms based on belief propagation

Abstract

Message-passing algorithms based on the Belief Propagation (BP) equations constitute a well-known distributed computational scheme. It is exact on tree-like graphical models and has also proven to be effective in many problems defined on graphs with loops (from inference to optimization, from signal processing to clustering). The BP-based scheme is fundamentally different from stochastic gradient descent (SGD), on which the current success of deep networks is based. In this paper, we present and adapt to mini-batch training on GPUs a family of BP-based message-passing algorithms with a reinforcement field that biases distributions towards locally entropic solutions. These algorithms are capable of training multi-layer neural networks with discrete weights and activations with performance comparable to SGD-inspired heuristics (BinaryNet) and are naturally well-adapted to continual learning. Furthermore, using these algorithms to estimate the marginals of the weights allows us to make approximate Bayesian predictions that have higher accuracy than point-wise solutions.

References in corpus (35)

• Overcoming catastrophic forgetting in neural networks
• Message Passing Algorithms for Compressed Sensing
• Expectation Propagation for approximate Bayesian inference
• Continual Learning Through Synaptic Intelligence
• Binarized Neural Networks
• An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks
• Visualizing the Loss Landscape of Neural Nets
• Probabilistic Backpropagation for Scalable Learning of Bayesian Neural Networks
• Statistical physics of inference: Thresholds and algorithms
• Bilinear Generalized Approximate Message Passing
• Optimal Errors and Phase Transitions in High-Dimensional Generalized Linear Models
• Loss Surfaces, Mode Connectivity, and Fast Ensembling of DNNs
• Fantastic Generalization Measures and Where to Find Them
• Unreasonable Effectiveness of Learning Neural Networks: From Accessible States and Robust Ensembles to Basic Algorithmic Schemes
• Learning by message-passing in networks of discrete synapses
• Entropy and mutual information in models of deep neural networks
• Subdominant Dense Clusters Allow for Simple Learning and High Computational Performance in Neural Networks with Discrete Synapses
• Efficient supervised learning in networks with binary synapses
• Modelling the influence of data structure on learning in neural networks: the hidden manifold model
• Phase transitions and sample complexity in Bayes-optimal matrix factorization
• Mean-field message-passing equations in the Hopfield model and its generalizations
• Shaping the learning landscape in neural networks around wide flat minima
• Synaptic metaplasticity in binarized neural networks
• Pruning Neural Networks at Initialization: Why are We Missing the Mark?
• Mean-field inference methods for neural networks
• Sparsely-Connected Neural Networks: Towards Efficient VLSI Implementation of Deep Neural Networks
• Neural Enhanced Belief Propagation on Factor Graphs
• Multi-Layer Generalized Linear Estimation
• Entropic gradient descent algorithms and wide flat minima
• Multi-Layer Bilinear Generalized Approximate Message Passing
• How Do Adam and Training Strategies Help BNNs Optimization?
• Perturbative construction of mean-field equations in extensive-rank matrix factorization and denoising
• Learning may need only a few bits of synaptic precision
• Belief Propagation Neural Networks
• Critical initialisation in continuous approximations of binary neural networks

Cited by in corpus (5)

• Perturbative construction of mean-field equations in extensive-rank matrix factorization and denoising
• Phase transitions in the mini-batch size for sparse and dense two-layer neural networks
• FlatNAS: optimizing Flatness in Neural Architecture Search for Out-of-Distribution Robustness
• Architecture-Aware Minimization (A${}^2$M): How to Find Flat Minima in Neural Architecture Search
• Training Multi-Layer Binary Neural Networks With Local Binary Error Signals