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

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