Exploring Categorization Strategies — ScienceDaily

Our mental ability to divide the complex world into categories makes our daily life much easier. But how do we categorize? What kind of stimulus properties are we evaluating? Researchers have gone a step further to answer these questions with the help of pigeons. They found that birds use different strategies to successfully learn categories. To collect data, the researchers used a new research method. To this end, they combined so-called virtual phylogenesis, in which artificial stimuli are generated by computers, with a machine learning approach, namely an automated assessment of the pecking behavior of birds.

Known colloquially as locker, learning by categorization often has a rather negative connotation in the eyes of the public. Yet the basic cognitive ability to categorize offers a significant advantage: it condenses the stream of objects and events in our environment based on commonalities and makes the knowledge we have accumulated usable for new experiences.

In science, the aspects of stimuli that determine categorization have long been the subject of contentious debate. The study conducted by the Bochum-based research team now offers insight into this question – through a research approach using computer-generated stimuli in combination with machine learning analysis of the pecking behavior of pigeons. “We specialize in working with these animals,” said Dr. Roland Pusch, lead author of the study. “Pigeons have a highly developed visual system and show excellent performance in behavioral tests. This makes them an excellent model system to address this issue.”

The specific pecking behavior of pigeons facilitates a detailed analysis

The biopsychologists trained the pigeons to distinguish digitally produced images on the screen and divide them into categories by pecking at the screen. “We precisely defined the properties of the image stimuli,” as Pusch describes the process. “Thanks to the so-called virtual phylogenesis, we created two families of objects with 20 members each on the computer. Based on its properties, each object clearly belonged to family X or family Y and could therefore be categorized accordingly by the animals.” “The main highlight of our research series was the specific pecking behavior of pigeons”, adds Professor Onur Güntürkün, project leader. “After training, pigeons use pecking to indicate whether an object belongs to a category or not. At the same time, they also mark exactly the place of the object that was decisive for their choice of categorization.”

Based on the automated recording, the researchers identified the locations on the objects the pigeons touched as they made their choices on the monitor. “The pecking behavior of individual animals was very consistent. This leads us to the conclusion that animals attach importance to very specific features of stimuli,” says Pusch. “Interestingly, despite identical behaviors, these preferences are different in each individual, that is, each pigeon has its own specific characteristics that it considers important in both families of objects. This suggests that learning categorization is not limited to a single learning strategy.”

According to Pusch and Güntürkün, the combination of virtual phylogenesis and machine learning approach offers a lot of potential for further research in the field of categorization learning. For example, the method opens the possibility of studying species-specific behavioral strategies in comparative experiments in addition to its sensory basis. Beyond behavioral analysis, the neural processes that trigger categorization learning in the brain could also be explored in detail.

Source of the story:

Material provided by Ruhr-University of Bochum. Note: Content may be edited for style and length.

Comments are closed.