A theoretical framework for end-to-end learning of deep neural networks with applications to robotics

Deep Learning (DL) systems are difficult to analyze and proving convergence of DL algorithms like backpropagation is an extremely challenging task as it is a highly non-convex and high-dimensional problem. When using DL algorithms in robotic systems, theoretical analysis of stability, convergence, a...

Full description

Saved in:
Bibliographic Details
Main Authors: Li, Sitan, Nguyen, Huu-Thiet, Cheah, Chien Chern
Other Authors: School of Electrical and Electronic Engineering
Format: Article
Language:English
Published: 2023
Subjects:
Online Access:https://hdl.handle.net/10356/168831
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-168831
record_format dspace
spelling sg-ntu-dr.10356-1688312023-06-23T15:40:38Z A theoretical framework for end-to-end learning of deep neural networks with applications to robotics Li, Sitan Nguyen, Huu-Thiet Cheah, Chien Chern School of Electrical and Electronic Engineering Engineering::Electrical and electronic engineering Deep Neural Networks Online Learning Deep Learning (DL) systems are difficult to analyze and proving convergence of DL algorithms like backpropagation is an extremely challenging task as it is a highly non-convex and high-dimensional problem. When using DL algorithms in robotic systems, theoretical analysis of stability, convergence, and robustness is a vital procedure as robots need to operate in a predictable manner to ensure safety. This paper presents the first unified End-to-End (E2E) learning framework that can be applied to both classification problems and real-time kinematic robot control tasks. In the proposed forward simultaneous learning method, the weights of all layers of the fully connected neural networks are updated concurrently. The proposed E2E learning framework uses an adjustable ReLU activation function so that convergence of the output error to a bound, which is dependent on the neural network approximation error, can be ensured. In particular, it is shown that the error between ideal output and estimated output of the network can converge to and stay in a certain bound, and this bound reduces to zero when the approximation error is zero. Therefore, the robustness of the learning system can be ensured even in the presence of the approximation error. Two case studies are done on classification tasks using MNIST and CIFAR10 datasets by using the proposed learning method, and the results show that the E2E learning method achieves comparable test accuracy as the gradient descent method and the main advantage is that convergence can be ensured during training. The framework is also implemented on a UR5e robot with unknown kinematic model, which is the first result on real-time E2E learning of deep neural networks for kinematic control of robots with guaranteed convergence. Ministry of Education (MOE) Published version This work was supported by the Ministry of Education (MOE) Singapore, Academic Research Fund (AcRF) Tier 1, under Grant RG65/22. 2023-06-20T02:15:11Z 2023-06-20T02:15:11Z 2023 Journal Article Li, S., Nguyen, H. & Cheah, C. C. (2023). A theoretical framework for end-to-end learning of deep neural networks with applications to robotics. IEEE Access, 11, 21992-22006. https://dx.doi.org/10.1109/ACCESS.2023.3249280 2169-3536 https://hdl.handle.net/10356/168831 10.1109/ACCESS.2023.3249280 2-s2.0-85149401135 11 21992 22006 en RG65/22 IEEE Access This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ application/pdf
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Engineering::Electrical and electronic engineering
Deep Neural Networks
Online Learning
spellingShingle Engineering::Electrical and electronic engineering
Deep Neural Networks
Online Learning
Li, Sitan
Nguyen, Huu-Thiet
Cheah, Chien Chern
A theoretical framework for end-to-end learning of deep neural networks with applications to robotics
description Deep Learning (DL) systems are difficult to analyze and proving convergence of DL algorithms like backpropagation is an extremely challenging task as it is a highly non-convex and high-dimensional problem. When using DL algorithms in robotic systems, theoretical analysis of stability, convergence, and robustness is a vital procedure as robots need to operate in a predictable manner to ensure safety. This paper presents the first unified End-to-End (E2E) learning framework that can be applied to both classification problems and real-time kinematic robot control tasks. In the proposed forward simultaneous learning method, the weights of all layers of the fully connected neural networks are updated concurrently. The proposed E2E learning framework uses an adjustable ReLU activation function so that convergence of the output error to a bound, which is dependent on the neural network approximation error, can be ensured. In particular, it is shown that the error between ideal output and estimated output of the network can converge to and stay in a certain bound, and this bound reduces to zero when the approximation error is zero. Therefore, the robustness of the learning system can be ensured even in the presence of the approximation error. Two case studies are done on classification tasks using MNIST and CIFAR10 datasets by using the proposed learning method, and the results show that the E2E learning method achieves comparable test accuracy as the gradient descent method and the main advantage is that convergence can be ensured during training. The framework is also implemented on a UR5e robot with unknown kinematic model, which is the first result on real-time E2E learning of deep neural networks for kinematic control of robots with guaranteed convergence.
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Li, Sitan
Nguyen, Huu-Thiet
Cheah, Chien Chern
format Article
author Li, Sitan
Nguyen, Huu-Thiet
Cheah, Chien Chern
author_sort Li, Sitan
title A theoretical framework for end-to-end learning of deep neural networks with applications to robotics
title_short A theoretical framework for end-to-end learning of deep neural networks with applications to robotics
title_full A theoretical framework for end-to-end learning of deep neural networks with applications to robotics
title_fullStr A theoretical framework for end-to-end learning of deep neural networks with applications to robotics
title_full_unstemmed A theoretical framework for end-to-end learning of deep neural networks with applications to robotics
title_sort theoretical framework for end-to-end learning of deep neural networks with applications to robotics
publishDate 2023
url https://hdl.handle.net/10356/168831
_version_ 1772825654783377408