What am I Searching for: Zero-shot Target Identity Inference in Visual Search

TL;DR

This paper introduces InferNet, a zero-shot neural model that infers search targets from error fixations in eye movement data, outperforming baseline models without object-specific training.

Contribution

The paper presents a novel zero-shot inference model, InferNet, that decodes search targets from error fixations using pre-trained CNN features, without task-specific training.

Findings

InferNet accurately predicts search targets from error fixations.

The model outperforms baseline null models in target inference.

Zero-shot approach eliminates need for object-specific training.

Abstract

Can we infer intentions from a person's actions? As an example problem, here we consider how to decipher what a person is searching for by decoding their eye movement behavior. We conducted two psychophysics experiments where we monitored eye movements while subjects searched for a target object. We defined the fixations falling on \textit{non-target} objects as "error fixations". Using those error fixations, we developed a model (InferNet) to infer what the target was. InferNet uses a pre-trained convolutional neural network to extract features from the error fixations and computes a similarity map between the error fixations and all locations across the search image. The model consolidates the similarity maps across layers and integrates these maps across all error fixations. InferNet successfully identifies the subject's goal and outperforms competitive null models, even without any object-specific training on the inference task.