[Rescorla & Wagner]
I wonder to what extent attentional theory is necessarily mutually exclusive with Pavlovian conditioning. Specifically, I imagine that its relevance may vary depending on the context of the learning, and also on the cognitive abilities of the learner — e.g., human adults, children, infants, or different kinds of animals.
[Lake et al.]
The article speaks about the core ingredients for building human-like intelligence — machines that learn and think like people. It is striking to me that there is nothing in the area of affective cognition/emotional intelligence/emotional regulation/empathy, or related abilities. These seem core to a big part of how and why people learn and think, since I believe our thinking and learning are ultimately motivated by trying to bring about some kind of benefit to ourselves, each other, and our communities. Even discovering some seemingly abstract knowledge brings about a surge of joy (and related hormones) because it has some value to humans (even if on an obvious level, that benefit is simply the discovery of some new abstract knowledge, e.g., in Mathematics). But the same can’t be said for machines (at this stage). So how can one just ignore this crucial difference in a paper like this one?