Algorithm Development in Neural Networks: Insights from the Streaming Parity Task
This provides insights into algorithm development in neural networks, addressing a fundamental problem in machine learning about extrapolation mechanisms, though it is incremental as it focuses on a specific case study.
The paper tackled the problem of how neural networks can generalize infinitely beyond their training data by studying recurrent neural networks on the streaming parity task, showing that with sufficient finite training, they achieve a phase transition to perfect infinite generalization.
Even when massively overparameterized, deep neural networks show a remarkable ability to generalize. Research on this phenomenon has focused on generalization within distribution, via smooth interpolation. Yet in some settings neural networks also learn to extrapolate to data far beyond the bounds of the original training set, sometimes even allowing for infinite generalization, implying that an algorithm capable of solving the task has been learned. Here we undertake a case study of the learning dynamics of recurrent neural networks (RNNs) trained on the streaming parity task in order to develop an effective theory of algorithm development. The streaming parity task is a simple but nonlinear task defined on sequences up to arbitrary length. We show that, with sufficient finite training experience, RNNs exhibit a phase transition to perfect infinite generalization. Using an effective theory for the representational dynamics, we find an implicit representational merger effect which can be interpreted as the construction of a finite automaton that reproduces the task. Overall, our results disclose one mechanism by which neural networks can generalize infinitely from finite training experience.