To use avian probabilistic inferences as a model system for testing: (i) whether cognition evolves in a convergent manner and (ii) the nature of the biases and errors seen in human probabilistic inference. To do this, we will use a new experimental framework, the signature-testing approach, to create a series of multi-dimensional intelligence tests focusing on core aspects of probabilistic inference and then present these tests to humans, kea parrots, and rats.
Kea parrots, rats and humans.
Our research is focused on comparing and understanding the similarities and differences between human, animal and machine minds. At the heart of our research program is an experimental framework we term the signature testing approach (Taylor et al. 2022). This uses the information processing errors, biases and other patterns agent exhibits to make inference about the content of different minds. This approach generates powerful intelligence tests that strongly constrain the possible cognition an agent is using and allows for the comparison of both biological and artificial intelligences. We also have a keen interest in understanding the role information processing errors and biases play in human decision-making. Our lab currently works with a wide range of study species, including the kea parrot, rats, and adult humans.
