Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/147041
Title: Improving machine learning methods for solving non-stationary conditions based on data availability, time urgency, and types of change
Authors: Goh, Chun Fan
Keywords: Engineering::Mechanical engineering::Robots
Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Issue Date: 2021
Publisher: Nanyang Technological University
Source: Goh, C. F. (2021). Improving machine learning methods for solving non-stationary conditions based on data availability, time urgency, and types of change. Doctoral thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/147041
Abstract: Supervised learning algorithms do not work well when the deployment condition is dissimilar to the training condition. Such non-stationary conditions include covariate shifts and concept shifts. Importance weighted learning (IWL) is used to handle a one-time covariate shift but not frequent shifts and concept shifts. While forgetting addresses concept shifts, it is wasteful in discarding previously learned models. To address these shortfalls, this thesis proposes looking into the three stages of supervised learning and devised pre-learning methods, which deal with data and feature selection; in-learning methods, which modify the learning process; and post-learning methods, which modify the prediction process to compensate for shifts in conditions. The first in-learning method is a transfer learning-based technique that utilizes a limited amount of test data to train further the prediction model pre-trained on general training data. This technique boosted the accuracy of a vocal emotion recognizer by 10%. For applications that require a timely response, we employed a post-learning strategy in the form of local learning. It handles multiple covariate shifts and improves the prediction accuracy in one vocal emotion recognition instance from 88.8% to 93.2%. Local learning also allows the use of feature augmentation to convert a more difficult concept-shift problem into an easier covariate-shift problem. The resulting controller outperforms PID controllers in water shooting control. When data are abundant, we leverage pre-learning methods such as condition-specific learning, to avoid non-stationary conditions altogether. Using this technique, we developed a semi-automatic snore labeling software that produces good accuracy (0.93 F1-score) and cuts labeling time from hours to minutes. Alternatively, we use deep learning methods to learn features that are robust to shifts. In our ablation study, we showed that features extracted from very deep networks and recurrent networks result in a more accurate and robust snore classification. With the advance of computer simulation, unlimited artificial data can be generated to better approximate and cover possible test conditions. We tested this idea in teaching a double-hull welding robot to climb down safely from a high wall through reinforcement learning and achieved a 90% success rate. Finally, from these applications, we distilled a method selection guideline based on data availability, time urgency, and type of shift.
URI: https://hdl.handle.net/10356/147041
DOI: 10.32657/10356/147041
Rights: This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).
Fulltext Permission: open
Fulltext Availability: With Fulltext
Appears in Collections:MAE Theses

Files in This Item:
File Description SizeFormat 
gcf_phd_thesis_ntu_v3.pdf20.28 MBAdobe PDFView/Open

Page view(s)

229
Updated on Feb 3, 2023

Download(s) 50

151
Updated on Feb 3, 2023

Google ScholarTM

Check

Altmetric


Plumx

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.