Please use this identifier to cite or link to this item:
https://hdl.handle.net/10356/155416
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Dang, Zhang | en_US |
dc.date.accessioned | 2022-02-23T02:19:50Z | - |
dc.date.available | 2022-02-23T02:19:50Z | - |
dc.date.issued | 2021 | - |
dc.identifier.citation | Dang, Z. (2021). Modified-LwF method for continual learning. Master's thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/155416 | en_US |
dc.identifier.uri | https://hdl.handle.net/10356/155416 | - |
dc.description.abstract | In this dissertation, we show that it is possible to overcome the catastrophic forgetting with several different methods. What is more important is that our method remembers old tasks better by combining the original learning without forgetting and elastic weight consolidation, which is the main contribution that both the merits of elastic weight consolidation and learning without forgetting are put into one method (Modified LwF). Besides, the upper bound joint training method, fine tune, EWC and original LwF methods are experimented by adding the new tasks one by one. In this procedure, the paths of the training in the algorithm will be focused more on. We finally finished all four tasks, and the size of the fourth task is far bigger than the previous three. | en_US |
dc.language.iso | en | en_US |
dc.publisher | Nanyang Technological University | en_US |
dc.subject | Engineering::Electrical and electronic engineering | en_US |
dc.title | Modified-LwF method for continual learning | en_US |
dc.type | Thesis-Master by Coursework | en_US |
dc.contributor.supervisor | Ponnuthurai Nagaratnam Suganthan | en_US |
dc.contributor.school | School of Electrical and Electronic Engineering | en_US |
dc.description.degree | Master of Science (Computer Control and Automation) | en_US |
dc.contributor.supervisoremail | EPNSugan@ntu.edu.sg | en_US |
item.grantfulltext | restricted | - |
item.fulltext | With Fulltext | - |
Appears in Collections: | EEE Theses |
Google ScholarTM
Check
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.