The last few years have witnessed amazing breakthroughs in machine intelligence, using systems that rely on neural networks as advocated in Parallel Distributed Processing (Rumelhart, McClelland et al, 1986). Yet in this talk, I will argue, in agreement with Lake et al. (2017) and others, that we still have a long way to go before any machine has truly captured human like cognitive and learning abilities. Unlike Lake et al., I will argue that we should seek the reasons for many of the amazing achievements of human intelligence not in built in biases or special purpose start-up software, but in a fuller appreciation of the roles of culture and experience. I will argue for a central role for culturally constructed formal systems as powerful tools that extend human abilities beyond what can be achieved without these resources. I will also argue for a central role of language-based instruction and explanation.

The discrete entities and explicit relations of symbolic systems make them transparent and easy to communicate. This contrasts with distributed systems, which tend to be opaque. This can lead us to pursue symbolic characterizations of human cognition. Symbolic interpretations can, however, oversimplify distributed systems. This is demonstrated in the developmental number cognition literature, where recent findings suggest a gradience of counting ability in children's learning. We take inspiration from these findings to explore the meaning of symbols in Recurrent Neural Networks (RNNs). We align recurrent neural representations with number symbols by causally intervening on the neural representations. We find that symbol-like representations of numbers do emerge in RNNs. We use this to inform the discussion on how neural systems represent quantity. We also show that the symbol-like representations evolve with learning, and continue to vary after the RNNs solve the task, demonstrating the graded nature of symbols in distributed systems.

Education has a profound impact on human cognition. People who have participated in education are better at solvingabstract reasoning tasks, can flexibly transfer knowledge across domains and are better at explaining their solutions.However, the properties of education that are responsible for these cognitive changes are poorly understood. We explorethe hypothesis that a structured education consisting of a cumulative, compositional curricular learning regime usingculturally constructed concepts and tools can account for many of these observations. In particular, we demonstrate thata connectionist model that learns to solve difficult analogical reasoning problems using a structured education is betterat knowledge reuse, while simultaneously providing explanations for solutions. We predict that premature progressionthrough a curriculum, before proficiency in a foundational stage has been established can fundamentally limit the potentialfor subsequent abstract reasoning performance or knowledge transfer ability.