ALife 2022 special session

Artificial Perception

Machines with Lifelike Failings

At ALife: July 18 - July 22, 2022

July 21st , 10:00~11:30 am CET (time converter)

Concept

The easiest way to investigate a complex system is to study its failures.

The easiest way to investigate a complex system is to study its failures. This is an accepted methodology in fields that deal with human complexity, such as neuroscience, psychology, or behavioral economics, as well as in fields dealing with artificial systems’ complexity, such as engineering or information security/cryptography. Fields that rely on synthesizing biological processes in artificial media can benefit from reproducing interesting failures in artificial systems, yet this approach is underused in both Artificial Life and Artificial Intelligence.

The goal of this special session is to use the affinity of ALife research for the unique and the unexpected, to highlight this approach where the study of failures provides immediate benefits: perception studies. In humans or other animals, perception is notoriously fallible. Perception is imprecise, and sometimes inaccurate in spectacular ways: auditory, tactile, or visual illusions. What can we learn from machines that fail like humans do?

Call for papers

This special session welcomes the following topics under the umbrella term of “Artificial Perception”:

Development of machines/algorithms that fall for the same perceptual traps as biological organisms (eg, fake memories, visual / auditory / tactile illusions...)
Research on the advantages of inaccurate perception (eg, work related to the Interface Theory of Perception, change blindness, attentional blindness, the cocktail party effect, visual filling-in...)
Research on the differences between machine perception and biological perception
Research on machines susceptible to priming effects (eg, the Stroop effect)
Surveys of biological failures that could make interesting targets for emulation in artificial systems (eg, illusions seen by different animals)

We invite contributions from theoretical or practical perspectives, from the fields of cognitive science, evolutionary dynamics, Artificial Intelligence, psychology, animal behavior, computational neuroscience, behavioral economics, robotics, etc.

Submit