Comparatives, Quantifiers, Proportions: A Multi-Task Model for the Learning of Quantities from Vision
This work addresses the challenge of modeling human-like non-symbolic quantity estimation from vision, which is incremental in nature.
The paper tackled the problem of jointly learning different quantification mechanisms from visual scenes using a multi-task model, showing that the higher-level proportional task becomes more accurate when lower-complexity tasks are available, and the model generalizes to unseen object combinations.
The present work investigates whether different quantification mechanisms (set comparison, vague quantification, and proportional estimation) can be jointly learned from visual scenes by a multi-task computational model. The motivation is that, in humans, these processes underlie the same cognitive, non-symbolic ability, which allows an automatic estimation and comparison of set magnitudes. We show that when information about lower-complexity tasks is available, the higher-level proportional task becomes more accurate than when performed in isolation. Moreover, the multi-task model is able to generalize to unseen combinations of target/non-target objects. Consistently with behavioral evidence showing the interference of absolute number in the proportional task, the multi-task model no longer works when asked to provide the number of target objects in the scene.