Page 136 |
Save page Remove page | Previous | 136 of 151 | Next |
|
small (250x250 max)
medium (500x500 max)
Large (1000x1000 max)
Extra Large
large ( > 500x500)
Full Resolution
All (PDF)
|
This page
All
|
6.2 Future Work
There is still a long path to traverse until robots will be able to operate in unstructured real world
settings. One of the key challenges is incremental adaptation using scarce data available to the
robot. Robots should be able to disentangle samples of useful data from the entire information flow,
efficiently learn complex mappings from new sources, switch in a timely way between sources
of information, and quickly adapt based on a few samples available. Our work advances the
state-of-the-art a few steps toward this goal and leads to several important future work directions.
Perceptual Novelty Detection and Data Attention Models. In the constantly changing condi-tions
of the real world the robot must be able to effectively distinguish i) relevant data for
state estimation, such as, low-noise data allowing high observability of states; and ii) novel
and task-relevant observations to incorporate only important experiences. This mitigates the
problem of catastrophic forgetting and allows better fitting of important mappings.
Effective Selection and Expansions of Priors. As the robot discovers novel observations, it
should not learn how to use them from scratch. Humans effectively re-use already learned
representations to bootstrap learning. There is a great value in finding algorithms that
together with the novelty detection models would enable to constantly grow and re-use
learned perceptual mappings and motor skills.
Better Models of World Representation. Besides the re-use of the proper priors for bootstrap-ping,
we also need to incorporate structures into model architectures that allow better sample
efficiency. Since humans tend to think in terms of objects and their relations, we believe that
such inductive biases should be reflected in the structures of perceptual systems.
121
Object Description
| Title | Data scarcity in robotics: leveraging structural priors and representation learning |
| Author | Molchanov, Artem |
| Author email | a.molchanov86@gmail.com;molchano@usc.edu |
| Degree | Doctor of Philosophy |
| Document type | Dissertation |
| Degree program | Computer Science |
| School | Viterbi School of Engineering |
| Date defended/completed | 2020-05-11 |
| Date submitted | 2020-08-11 |
| Date approved | 2020-08-11 |
| Restricted until | 2020-08-11 |
| Date published | 2020-08-11 |
| Advisor (committee chair) | Sukhatme, Gaurav Suhas |
| Advisor (committee member) |
Ayanian, Nora Culbertson, Heather Gupta, Satyandra K. |
| Abstract | Recent advances in Artificial Intelligence have benefited significantly from access to large pools of data accompanied in many cases by labels, ground truth values, or perfect demonstrations. In robotics, however, such data are scarce or absent completely. Overcoming this issue is a major barrier to move robots from structured laboratory settings to the unstructured real world. In this dissertation, by leveraging structural priors and representation learning, we provide several solutions when data required to operate robotics systems is scarce or absent. ❧ In the first part of this dissertation we study sensory feedback scarcity. We show how to use high-dimensional alternative sensory modalities to extract data when primary sensory sources are absent. In a robot grasping setting, we address the problem of contact localization and solve it using multi-modal tactile feedback as the alternative source of information. We leverage multiple tactile modalities provided by electrodes and hydro-acoustic sensors to structure the problem as spatio-temporal inference. We employ the representational power of neural networks to acquire the complex mapping between tactile sensors and the contact locations. We also investigate scarce feedback due to the high cost of measurements. We study this problem in a challenging field robotics setting where multiple severely underactuated aquatic vehicles need to be coordinated. We show how to leverage collaboration among the vehicles and the spatio-temporal smoothness of the ocean currents as a prior to densify feedback about ocean currents in order to acquire better controllability. ❧ In the second part of this dissertation, we investigate scarcity of the data related to the desired task. We develop a method to efficiently leverage simulated dynamics priors to perform sim-to-real transfer of a control policy when no data about the target system is available. We investigate this problem in the scenario of sim-to-real transfer of low-level stabilizing quadrotor control policies. We demonstrate that we can learn robust policies in simulation and transfer them to the real system while acquiring no samples from the real quadrotor. Finally, we consider the general problem of learning a model with a very limited number of samples using meta-learned losses. We show how such losses can encode a prior structure about families of tasks to create well-behaved loss landscapes for efficient model optimization. We demonstrate the efficiency of our approach for learning policies and dynamics models in multiple robotics settings. |
| Keyword | robotics; machine learning; artificial intelligence |
| Language | English |
| Part of collection | University of Southern California dissertations and theses |
| Publisher (of the original version) | University of Southern California |
| Place of publication (of the original version) | Los Angeles, California |
| Publisher (of the digital version) | University of Southern California. Libraries |
| Provenance | Electronically uploaded by the author |
| Type | texts |
| Legacy record ID | usctheses-m |
| Contributing entity | University of Southern California |
| Rights | Molchanov, Artem |
| Physical access | The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given. |
| Repository name | University of Southern California Digital Library |
| Repository address | USC Digital Library, University of Southern California, University Park Campus MC 7002, 106 University Village, Los Angeles, California 90089-7002, USA |
| Repository email | cisadmin@lib.usc.edu |
| Filename | etd-MolchanovA-8923.pdf |
| Archival file | Volume13/etd-MolchanovA-8923.pdf |
Description
| Title | Page 136 |
| Full text | 6.2 Future Work There is still a long path to traverse until robots will be able to operate in unstructured real world settings. One of the key challenges is incremental adaptation using scarce data available to the robot. Robots should be able to disentangle samples of useful data from the entire information flow, efficiently learn complex mappings from new sources, switch in a timely way between sources of information, and quickly adapt based on a few samples available. Our work advances the state-of-the-art a few steps toward this goal and leads to several important future work directions. Perceptual Novelty Detection and Data Attention Models. In the constantly changing condi-tions of the real world the robot must be able to effectively distinguish i) relevant data for state estimation, such as, low-noise data allowing high observability of states; and ii) novel and task-relevant observations to incorporate only important experiences. This mitigates the problem of catastrophic forgetting and allows better fitting of important mappings. Effective Selection and Expansions of Priors. As the robot discovers novel observations, it should not learn how to use them from scratch. Humans effectively re-use already learned representations to bootstrap learning. There is a great value in finding algorithms that together with the novelty detection models would enable to constantly grow and re-use learned perceptual mappings and motor skills. Better Models of World Representation. Besides the re-use of the proper priors for bootstrap-ping, we also need to incorporate structures into model architectures that allow better sample efficiency. Since humans tend to think in terms of objects and their relations, we believe that such inductive biases should be reflected in the structures of perceptual systems. 121 |
