Few-shot and transfer learning with manifold distributed datasets

Sayed Waleed Qayyumi, Laurence A.F. Park, Oliver Obst

Research output: Chapter in Book / Conference PaperConference Paperpeer-review

Abstract

A manifold distributed dataset with limited labels makes it difficult to train a high-mean accuracy classifier. Transfer learning is beneficial in such circumstances. For transfer learning to succeed, the target and base datasets should have a similar manifold structure. A novel method is presented in this paper for determining the similarity between two manifold structures. To determine whether target and base datasets have similar manifolds and are suitable for transfer learning, this method can be used. A novel few-shot algorithm is then presented that uses transfer learning to classify manifold distributed datasets with a limited number of labels. Using the base and target datasets, the manifold structure and its relevant label distribution are learned. Using this information in combination with the few labels and unlabeled data from the target dataset, we can develop a classifier with high mean accuracy.
Original languageEnglish
Title of host publicationData Science and Machine Learning
Subtitle of host publication21st Australasian Conference, AusDM 2023, Auckland, New Zealand, December 11-13 2023
EditorsDiana Benavides-Prado, Sarah Erfani, Philippe Fournier-Viger, Yee Ling Boo, Yun Sing Koh
Place of PublicationSingapore
PublisherSpringer Nature Singapore
Pages137-149
Number of pages13
ISBN (Electronic)9789819986965
ISBN (Print)9789819986958
DOIs
Publication statusPublished - 2024
Event21st Australasian Conference on Data Science and Machine Learning, AusDM 2023 - Auckland, New Zealand
Duration: 11 Dec 202313 Dec 2023

Publication series

NameCommunications in Computer and Information Science
Volume1943
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference21st Australasian Conference on Data Science and Machine Learning, AusDM 2023
Country/TerritoryNew Zealand
CityAuckland
Period11/12/2313/12/23

Keywords

  • Few-shot learning
  • manifold distributed datasets
  • measuring similarity
  • Transfer learning

Fingerprint

Dive into the research topics of 'Few-shot and transfer learning with manifold distributed datasets'. Together they form a unique fingerprint.

Cite this