2016/04/30 by Dinesh Jayaraman, Jayaraman, Dinesh, Kristen Grauman +1 · 2 citations
Computer Science · #Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Human Pose and Action Recognition #Machine Learning (cs.LG) #Multimodal Machine Learning Applications #Robotics (cs.RO)
paper · pdf · doi:10.48550/arxiv.1605.00164
openalex publication_date 2016/04/30 · openalex created_date 2022/10/05 · openalex updated_date 2026/07/28
Visual recognition systems mounted on autonomous moving agents face the\nchallenge of unconstrained data, but simultaneously have the opportunity to\nimprove their performance by moving to acquire new views of test data. In this\nwork, we first show how a recurrent neural network-based system may be trained\nto perform end-to-end learning of motion policies suited for this "active\nrecognition" setting. Further, we hypothesize that active vision requires an\nagent to have the capacity to reason about the effects of its motions on its\nview of the world. To verify this hypothesis, we attempt to induce this\ncapacity in our active recognition pipeline, by simultaneously learning to\nforecast the effects of the agent's motions on its internal representation of\nthe environment conditional on all past views. Results across two challenging\ndatasets confirm both that our end-to-end system successfully learns meaningful\npolicies for active category recognition, and that "learning to look ahead"\nfurther boosts recognition performance.\n