A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Filters
Zero-Shot Visual Imitation
2018
2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
Preserved Fulltext
These skills can then be used to imitate the visual demonstration provided by the expert [15] . ...
K d R p x C P Z D I g C z g T U N d M c m r E E E g Y c G s H g f O w 3 7 k E q F o l b P Y z B D 0 l P s C 6 j R B u p X S h 7 A f S Y S C k I D X J k l 4 J D f J V w z Y 6 U h h We call our method zero-shot ...
We compare to their method in both visual navigation and manipulation. (2) GSP-NoPrevAction-NoFwdConst is the ablation of our recurrent GSP without previous action history and without forward consistency ...
doi:10.1109/cvprw.2018.00278
dblp:conf/cvpr/PathakMLACSSMED18
fatcat:fkbiut3ttbgujfaswf3ar2a6nm
Zero-Shot Visual Imitation
[article]
2018
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We evaluate our zero-shot imitator in two real-world settings: complex rope manipulation with a Baxter robot and navigation in previously unseen office environments with a TurtleBot. ...
Our method is 'zero-shot' in the sense that the agent never has access to expert actions during training or for the task demonstration at inference. ...
We evaluate our zero-shot imitator on real-world robots for rope manipulation tasks using a Baxter and office navigation using a TurtleBot. ...
arXiv:1804.08606v1
fatcat:dgrsryxhqvbfhiqmnmjkyyakqi
Zero-shot Imitation Learning from Demonstrations for Legged Robot Visual Navigation
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Here, we propose a zero-shot imitation learning approach for training a visual navigation policy on legged robots from human (third-person perspective) demonstrations, enabling high-quality navigation ...
Imitation learning is a popular approach for training visual navigation policies. ...
METHOD This section introduces a zero-shot imitation learning framework for visual navigation of a legged robot. ...
arXiv:1909.12971v2
fatcat:5yedpru2wnchxg3zwagupq262e
Detection and Captioning with Unseen Object Classes
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Our experiments show that the proposed zero-shot detection model obtains state-of-the-art performance on the MS-COCO dataset and the zero-shot captioning approach yields promising results. ...
For this problem, we propose a detection-driven approach based on a generalized zero-shot detection model and a template-based sentence generation model. ...
Noticeably, the performance gap between true zero-shot and (visually) supervised partial zero-shot captioning is larger in terms of the Avg. F1 metric. ...
arXiv:2108.06165v1
fatcat:k4riatvu6vcknggyiwq2kbjcze
Visual Goal-Directed Meta-Learning with Contextual Planning Networks
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
We evaluate CPN along with several other approaches adapted for zero-shot goal-directed meta-learning. ...
We adapted the metaworld benchmark tasks to create 24 zero-shot meta-learning from visual demonstration tasks for evaluation. ...
Single-shot imitation learning via images has also been explored [22] , [23] , [1] , [24] . ...
arXiv:2111.09908v1
fatcat:tzl2dnkv5ne57nleq3guwwv77m
Gated-Attention Architectures for Task-Oriented Language Grounding
2018
PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
To perform tasks specified by natural language instructions, autonomous agents need to extract semantically meaningful representations of language and map it to visual elements and actions in the environment ...
The proposed model combines the image and text representations using a Gated-Attention mechanism and learns a policy to execute the natural language instruction using standard reinforcement and imitation ...
Multitask and Zero-Shot task generalization, across three modes of difficulty. ...
doi:10.1609/aaai.v32i1.11832
fatcat:4mekny7girh2hghlcldumbl4ni
Gated-Attention Architectures for Task-Oriented Language Grounding
[article]
2018
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
Other Versions
To perform tasks specified by natural language instructions, autonomous agents need to extract semantically meaningful representations of language and map it to visual elements and actions in the environment ...
The proposed model combines the image and text representations using a Gated-Attention mechanism and learns a policy to execute the natural language instruction using standard reinforcement and imitation ...
Multitask and Zero-Shot task generalization, across three modes of difficulty. ...
arXiv:1706.07230v2
fatcat:pckcwi6gbbaqzoiesionmcsrou
Towards More Generalizable One-shot Visual Imitation Learning
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
We then study the multi-task setting, where multi-task training is followed by (i) one-shot imitation on variations within the training tasks, (ii) one-shot imitation on new tasks, and (iii) fine-tuning ...
For consistency and comparison purposes, we first train and evaluate single-task agents (as done in prior few-shot imitation work). ...
One-shot imitation learning. ...
arXiv:2110.13423v2
fatcat:dxstakcokjgghabxh65y75iweq
BC-Z: Zero-Shot Task Generalization with Robotic Imitation Learning
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
We approach the challenge from an imitation learning perspective, aiming to study how scaling and broadening the data collected can facilitate such generalization. ...
To that end, we develop an interactive and flexible imitation learning system that can learn from both demonstrations and interventions and can be conditioned on different forms of information that convey ...
Daniel Kappler built the data annotation visualizer. Corey Lynch advised Frederik's internship and gave pointers on language models. Sergey Levine and Chelsea Finn supervised the project. ...
arXiv:2202.02005v1
fatcat:v2hr2vhlsjhubbngedinluefge
Visual Adversarial Imitation Learning using Variational Models
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Towards addressing these challenges, we develop a variational model-based adversarial imitation learning (V-MAIL) algorithm. ...
In contrast, providing visual demonstrations of desired behaviors often presents an easier and more natural way to teach agents. ...
Algorithm 2 Zero-Shot Transfer with V-MAIL prior model-free imitation learning approaches, and behavior cloning on five visual imitation tasks. ...
arXiv:2107.08829v2
fatcat:lp5atolewne37kcflwhfxfc76u
Joint Hypergraph Learning using feature fusion for Image Retrieval
2020
VOLUME-8 ISSUE-10, AUGUST 2019, REGULAR ISSUE
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
In this exploration field, label data and various visual highlights have been explored. Be that as it may, most existing strategies utilize these visual includes independently or successively. ...
In this paper, we propose a worldwide and neighborhood visual highlights combination way to deal with get familiar with the significance of pictures by hypergraph approach. ...
Subsequently, a little calculations recreation in imitation of mix various visual highlights in imitation of enhance the picture excerpt exactness. ...
doi:10.35940/ijitee.h6474.0891020
fatcat:byco263w75av7apwwwxxoluwvi
Generalization Through Hand-Eye Coordination: An Action Space for Learning Spatially-Invariant Visuomotor Control
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Through a set of challenging multi-stage manipulation tasks, we show that a visuomotor policy equipped with HAN is able to inherit the key spatial invariance property of hand-eye coordination and achieve zero-shot ...
Imitation Learning (IL) is an effective framework to learn visuomotor skills from offline demonstration data. ...
EXPERIMENTS In this section, we seek to answer the following questions: (1) Does including HAN in a deep imitation learning pipeline improve the task performance and zero-shot generalization ability? ...
arXiv:2103.00375v2
fatcat:iwzmwsmbmjfclekqf4xyejxvza
Learning from Observation-Only Demonstration for Task-Oriented Language Grounding via Self-Examination
2019
Neural Information Processing Systems
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Combining imitation with natural language instruction promises to further make imitation learning more flexible and useful in real-world applications. ...
Imitation learning is an effective method for learning a control policy from expert demonstrations. ...
Detailed Analysis Zero-shot Generalization: To investigate the generability of visual-language grounding, we evaluate under zero-shot setting where new combinations of attribute-object pairs are unseen ...
dblp:conf/nips/FuTKK19
fatcat:xtsyvbosz5eurdxxodasyhvqb4
One-Shot Visual Imitation Learning via Meta-Learning
[article]
2017
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Unlike prior methods for one-shot imitation, our method can scale to raw pixel inputs and requires data from significantly fewer prior tasks for effective learning of new skills. ...
Our experiments on both simulated and real robot platforms demonstrate the ability to learn new tasks, end-to-end, from a single visual demonstration. ...
Discussion and Future Work We proposed a method for one-shot visual imitation learning that can learn to perform tasks using visual inputs from just a single demonstration. ...
arXiv:1709.04905v1
fatcat:vz5ykqgh3zg3nae6kzv77h5v2e
FlowControl: Optical Flow Based Visual Servoing
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We present a practical method for realizing one-shot imitation for manipulation tasks, exploiting modern learning-based optical flow to perform real-time visual servoing. ...
One-shot imitation is the vision of robot programming from a single demonstration, rather than by tedious construction of computer code. ...
Few-shot imitation from videos is an appealing alternative to overcome this problem, as videos typically capture all task-relevant information. ...
arXiv:2007.00291v1
fatcat:7whjrzathfdxfmpcjhmh2usdr4
« Previous
Showing results 1 — 15 out of 7,117 results