Learning credit assignment
This work addresses the lack of transparency in deep learning for researchers and practitioners, offering a theoretical explanation for credit assignment, though it is incremental as it builds on existing mean-field approaches.
The authors tackled the problem of understanding hierarchical credit assignment in deep learning by proposing a mean-field learning model that assumes an ensemble of sub-networks is trained for classification. Their model reveals that synaptic connections can be categorized into three types—very important, unimportant, and variable—and predicts an ensemble of sub-networks that achieve the same task, providing insights into deep learning's macroscopic behavior.
Deep learning has achieved impressive prediction accuracies in a variety of scientific and industrial domains. However, the nested non-linear feature of deep learning makes the learning highly non-transparent, i.e., it is still unknown how the learning coordinates a huge number of parameters to achieve a decision making. To explain this hierarchical credit assignment, we propose a mean-field learning model by assuming that an ensemble of sub-networks, rather than a single network, are trained for a classification task. Surprisingly, our model reveals that apart from some deterministic synaptic weights connecting two neurons at neighboring layers, there exist a large number of connections that can be absent, and other connections can allow for a broad distribution of their weight values. Therefore, synaptic connections can be classified into three categories: very important ones, unimportant ones, and those of variability that may partially encode nuisance factors. Therefore, our model learns the credit assignment leading to the decision, and predicts an ensemble of sub-networks that can accomplish the same task, thereby providing insights toward understanding the macroscopic behavior of deep learning through the lens of distinct roles of synaptic weights.