From behavioral measurements to models that behave

Organisers: Warren Mansell, School of Health Sciences, University of Manchester, Manchester, UK and Alex Gomez-Marin, Institute of Neuroscience, CSIC-UMH, Alicante, Spain

Short name (in EasyChair): Models (S3)


Measuring behavior is not enough to understand it. For instance, the behavior of a fan controls its output, while a thermostat controls its input using its behavior. This distinction is crucial, since one must then acknowledge that behavior in humans, animals and robots, to be adaptive, has to be a closed-loop process. One can then discover the perceptual variables that the organism controls, which is different from how it responds to certain stimuli. To some extent, building is more challenging than intervening or fixing: grounded in (but not restricted to) behavioral measurements, a model that behaves can be more insightful than a model of behavior. Complementary to classification and extrapolation, we seek to discuss relevant methods and techniques to build and test generative models of behavior whose parameters are derived from empirical data.