The human cognitive system exhibits remarkable flexibility and generalization
capabilities, partly due to its ability to form low-dimensional, compositional
representations of the environment. In contrast, standard neural network
architectures often struggle with abstract reasoning tasks, overfitting, and
requiring extensive data for training. This paper investigates the impact of
the relational bottleneck -- a mechanism that focuses processing on relations
among inputs -- on the learning of factorized representations conducive to
compositional coding and the attendant flexibility of processing. We
demonstrate that such a bottleneck not only improves generalization and
learning efficiency, but also aligns network performance with human-like
behavioral biases. Networks trained with the relational bottleneck developed
orthogonal representations of feature dimensions latent in the dataset,
reflecting the factorized structure thought to underlie human cognitive
flexibility. Moreover, the relational network mimics human biases towards
regularity without pre-specified symbolic primitives, suggesting that the
bottleneck fosters the emergence of abstract representations that confer
flexibility akin to symbols.

研究关系瓶颈对因式表示学习和处理灵活性的影响，并证明该瓶颈不仅改善了泛化和学习效率，而且与人类行为偏差相一致，从而促进了抽象表示的产生，使处理灵活性类似于符号。