Gazing at Rewards: Eye Movements as a Lens into Human and AI Decision-Making in Hybrid Visual Foraging

TL;DR

This paper investigates how humans and AI models make reward-based decisions during visual foraging, revealing that eye movements reflect value-driven choices and introducing a transformer-based model that mimics human foraging behavior.

Contribution

It introduces a reinforcement learning-based transformer model that replicates human eye movement patterns and decision-making in a complex visual foraging task.

Findings

Humans focus on higher-value targets and outperform chance in reward collection.

The Visual Forager model achieves human-like foraging efficiency and eye movement patterns.

The model generalizes well to novel targets and varying task conditions.

Abstract

Imagine searching a collection of coins for quarters ($0.25$), dimes ($0.10$), nickels ($0.05$), and pennies ($0.01$)-a hybrid foraging task where observers look for multiple instances of multiple target types. In such tasks, how do target values and their prevalence influence foraging and eye movement behaviors (e.g., should you prioritize rare quarters or common nickels)? To explore this, we conducted human psychophysics experiments, revealing that humans are proficient reward foragers. Their eye fixations are drawn to regions with higher average rewards, fixation durations are longer on more valuable targets, and their cumulative rewards exceed chance, approaching the upper bound of optimal foragers. To probe these decision-making processes of humans, we developed a transformer-based Visual Forager (VF) model trained via reinforcement learning. Our VF model takes a series of targets, their corresponding values, and the search image as inputs, processes the images using foveated vision, and produces a sequence of eye movements along with decisions on whether to collect each fixated item. Our model outperforms all baselines, achieves cumulative rewards comparable to those of humans, and approximates human foraging behavior in eye movements and foraging biases within time-limited environments. Furthermore, stress tests on out-of-distribution tasks with novel targets, unseen values, and varying set sizes demonstrate the VF model's effective generalization. Our work offers valuable insights into the relationship between eye movements and decision-making, with our model serving as a powerful tool for further exploration of this connection. All data, code, and models are available at https://github.com/ZhangLab-DeepNeuroCogLab/visual-forager.