SONG: Self-Organizing Neural Graphs
This addresses the problem of interpretable and efficient neural network design for machine learning practitioners, though it appears incremental as it builds on decision graphs with a new training method.
The paper tackled the lack of efficient gradient-based training for decision graphs in deep learning by introducing Self-Organizing Neural Graphs (SONG), a paradigm based on Markov processes, and showed that it performs on par or better than existing decision models on datasets like MNIST and CIFAR.
Recent years have seen a surge in research on deep interpretable neural networks with decision trees as one of the most commonly incorporated tools. There are at least three advantages of using decision trees over logistic regression classification models: they are easy to interpret since they are based on binary decisions, they can make decisions faster, and they provide a hierarchy of classes. However, one of the well-known drawbacks of decision trees, as compared to decision graphs, is that decision trees cannot reuse the decision nodes. Nevertheless, decision graphs were not commonly used in deep learning due to the lack of efficient gradient-based training techniques. In this paper, we fill this gap and provide a general paradigm based on Markov processes, which allows for efficient training of the special type of decision graphs, which we call Self-Organizing Neural Graphs (SONG). We provide an extensive theoretical study of SONG, complemented by experiments conducted on Letter, Connect4, MNIST, CIFAR, and TinyImageNet datasets, showing that our method performs on par or better than existing decision models.