Hutter, M. | bioRxiv Science

bioRxiv2024

When monkeys meet an ANYmal robot in the wild

Animal-robot interaction studies have been of increasing interest in research, but most of these studies have involved robots interacting with insects, birds, and frogs in laboratory settings. To date, only two studies used non-human primates and no behavioral study has tested the social integration of a robot in a group of wild primates. To fill this gap, we studied the interactions between the quadruped ANYmal robot and a group of 37 wild vervet monkeys in South Africa. The ANYmal robot is a remote-controlled sheep-sized robot with an open box of food on its back. We gradually introduced the robot to the monkeys following five different steps over 6 days for a total exposition time of about 10h. The monkeys habituated to the robot very quickly with six individuals eating the food in the robots box from the second day. A few individuals, mostly juveniles, emitted alarm calls towards the robot. In total, seven individuals, high rankers, spent time near the robot and 21 monkeys approached the robot from a greater distance. High rankers displayed significantly more vigilant and self-centered behaviors and they, with females and juveniles, ate more food in the robots box compared to low rankers, males, and adults conversely. This study offers exciting perspectives on the phenomena of social acceptance of machines in mammalian societies and the automation of field data collection.

animal behavior and cognition↗

bioRxiv2021

A rapid and efficient learning rule for biological neural circuits

The dominant view in neuroscience is that changes in synaptic weights underlie learning. It is unclear, however, how the brain is able to determine which synapses should change, and by how much. This uncertainty stands in sharp contrast to deep learning, where changes in weights are explicitly engineered to optimize performance. However, the main tool for that, backpropagation, has two problems. One is neuro-science related: it is not biologically plausible. The other is inherent: networks trained with this rule tend to forget old tasks when learning new ones. Here we introduce the Dendritic Gated Network (DGN), a variant of the Gated Linear Network, which offers a biologically plausible alternative to backpropagation. DGNs combine dendritic gating (whereby interneurons target dendrites to shape neuronal responses) with local learning rules to yield provably efficient performance. They are significantly more data efficient than conventional artificial networks, and are highly resistant to forgetting. Consequently, they perform well on a variety of tasks, in some cases better than backpropagation. Importantly, DGNs have structural and functional similarities to the cerebellum, a link that we strengthen by using in vivo two-photon calcium imaging to show that single interneurons suppress activity in individual dendritic branches of Purkinje cells, a key feature of the model. Thus, DGNs leverage targeted dendritic inhibition and local learning - two features ubiquitous in the brain - to achieve fast and efficient learning.

neuroscience↗