The Deep Equilibrium Algorithmic Reasoner
This work addresses the challenge of efficiently training neural networks to execute classical algorithms, which could benefit researchers and practitioners in machine learning and algorithmic reasoning.
The authors tackled the problem of neural algorithmic reasoning by proposing a method that directly finds the equilibrium of algorithms using graph neural networks, eliminating the need for recurrent architectures that align iterations with algorithm steps, and empirically validated this approach.
Recent work on neural algorithmic reasoning has demonstrated that graph neural networks (GNNs) could learn to execute classical algorithms. Doing so, however, has always used a recurrent architecture, where each iteration of the GNN aligns with an algorithm's iteration. Since an algorithm's solution is often an equilibrium, we conjecture and empirically validate that one can train a network to solve algorithmic problems by directly finding the equilibrium. Note that this does not require matching each GNN iteration with a step of the algorithm.