Better Set Representations For Relational Reasoning
This addresses a fundamental limitation in relational reasoning models for AI applications, though it is incremental as it builds on existing approaches.
The paper tackled the problem of existing neural network approaches for relational reasoning not respecting set permutational invariance, leading to representational limitations, and proposed a Set Refiner Network (SRN) module that resulted in substantial gains in prediction performance and robustness on relational reasoning tasks.
Incorporating relational reasoning into neural networks has greatly expanded their capabilities and scope. One defining trait of relational reasoning is that it operates on a set of entities, as opposed to standard vector representations. Existing end-to-end approaches typically extract entities from inputs by directly interpreting the latent feature representations as a set. We show that these approaches do not respect set permutational invariance and thus have fundamental representational limitations. To resolve this limitation, we propose a simple and general network module called a Set Refiner Network (SRN). We first use synthetic image experiments to demonstrate how our approach effectively decomposes objects without explicit supervision. Then, we insert our module into existing relational reasoning models and show that respecting set invariance leads to substantial gains in prediction performance and robustness on several relational reasoning tasks.