Weight Priors for Learning Identity Relations
This addresses a long-standing issue in neural network learning for AI researchers, offering a potential solution for improving systematic reasoning, though it appears incremental as an extension of an existing method.
The paper tackles the problem of neural networks failing to learn identity relations and generalize to unseen data by extending the Relation Based Pattern approach with Bayesian weight priors, resulting in perfect generalization in experiments without hindering general learning.
Learning abstract and systematic relations has been an open issue in neural network learning for over 30 years. It has been shown recently that neural networks do not learn relations based on identity and are unable to generalize well to unseen data. The Relation Based Pattern (RBP) approach has been proposed as a solution for this problem. In this work, we extend RBP by realizing it as a Bayesian prior on network weights to model the identity relations. This weight prior leads to a modified regularization term in otherwise standard network learning. In our experiments, we show that the Bayesian weight priors lead to perfect generalization when learning identity based relations and do not impede general neural network learning. We believe that the approach of creating an inductive bias with weight priors can be extended easily to other forms of relations and will be beneficial for many other learning tasks.