Visual semantic planning using deep successor representations
A crucial capability of real-world intelligent agents is their ability to plan a sequence of
actions to achieve their goals in the visual world. In this work, we address the problem of …
actions to achieve their goals in the visual world. In this work, we address the problem of …
An exploration of embodied visual exploration
SK Ramakrishnan, D Jayaraman… - International Journal of …, 2021 - Springer
Embodied computer vision considers perception for robots in novel, unstructured
environments. Of particular importance is the embodied visual exploration problem: how …
environments. Of particular importance is the embodied visual exploration problem: how …
Learning to look around: Intelligently exploring unseen environments for unknown tasks
D Jayaraman, K Grauman - Proceedings of the IEEE …, 2018 - openaccess.thecvf.com
It is common to implicitly assume access to intelligently captured inputs (eg, photos from a
human photographer), yet autonomously capturing good observations is itself a major …
human photographer), yet autonomously capturing good observations is itself a major …
Embodied amodal recognition: Learning to move to perceive objects
Passive visual systems typically fail to recognize objects in the amodal setting where they
are heavily occluded. In contrast, humans and other embodied agents have the ability to …
are heavily occluded. In contrast, humans and other embodied agents have the ability to …
Look-ahead before you leap: end-to-end active recognition by forecasting the effect of motion
D Jayaraman, K Grauman - Computer Vision–ECCV 2016: 14th European …, 2016 - Springer
Visual recognition systems mounted on autonomous moving agents face the challenge of
unconstrained data, but simultaneously have the opportunity to improve their performance …
unconstrained data, but simultaneously have the opportunity to improve their performance …
Reinforcement learning of active vision for manipulating objects under occlusions
R Cheng, A Agarwal… - Conference on Robot …, 2018 - proceedings.mlr.press
We consider artificial agents that learn to jointly control their gripper and camera in order to
reinforcement learn manipulation policies in the presence of occlusions from distractor …
reinforcement learn manipulation policies in the presence of occlusions from distractor …
Sharing user IoT devices in the cloud
Y Benazzouz, C Munilla, O Günalp… - 2014 IEEE world …, 2014 - ieeexplore.ieee.org
Internet of Things (IoT) is the set of technologies that can interconnect anything, from daily
life objects to more sophisticated networked devices. The IoT paradigm is constantly …
life objects to more sophisticated networked devices. The IoT paradigm is constantly …
End-to-end policy learning for active visual categorization
D Jayaraman, K Grauman - IEEE transactions on pattern …, 2018 - ieeexplore.ieee.org
Visual recognition systems mounted on autonomous moving agents face the challenge of
unconstrained data, but simultaneously have the opportunity to improve their performance …
unconstrained data, but simultaneously have the opportunity to improve their performance …
Robotic material perception using active multimodal fusion
Robotic material perception is an extremely important but challenging problem for industrial
intelligence. The main difficulties come from the fact that the material properties are difficult …
intelligence. The main difficulties come from the fact that the material properties are difficult …
Extreme trust region policy optimization for active object recognition
H Liu, Y Wu, F Sun - IEEE transactions on neural networks and …, 2018 - ieeexplore.ieee.org
In this brief, we develop a deep reinforcement learning method to actively recognize objects
by choosing a sequence of actions for an active camera that helps to discriminate between …
by choosing a sequence of actions for an active camera that helps to discriminate between …