Inductive Linear Probing for Few-Shot Node Classification

Hirthik Mathavan; Zhen Tan; Nivedh Mudiam; Huan Liu

doi:10.1007/978-3-031-43129-6_27

Inductive Linear Probing for Few-Shot Node Classification

Hirthik Mathavan, Zhen Tan, Nivedh Mudiam, Huan Liu

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Meta-learning has emerged as a powerful training strategy for few-shot node classification, demonstrating its effectiveness in the transductive setting. However, the existing literature predominantly focuses on transductive few-shot node classification, neglecting the widely studied inductive setting in the broader few-shot learning community. This oversight limits our comprehensive understanding of the performance of meta-learning based methods on graph data. In this work, we conduct an empirical study to highlight the limitations of current frameworks in the inductive few-shot node classification setting. Additionally, we propose applying a competitive baseline approach specifically tailored for inductive few-shot node classification tasks. We hope our work can provide a new path forward to better understand how the meta-learning paradigm works in the graph domain.

Original language	English (US)
Title of host publication	Social, Cultural, and Behavioral Modeling - 16th International Conference, SBP-BRiMS 2023, Proceedings
Editors	Robert Thomson, Samer Al-khateeb, Annetta Burger, Patrick Park, Aryn A. Pyke
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	274-284
Number of pages	11
ISBN (Print)	9783031431289
DOIs	https://doi.org/10.1007/978-3-031-43129-6_27
State	Published - 2023
Event	16th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation, SBP-BRiMS 2023 - Pittsburgh, United States Duration: Sep 20 2023 → Sep 22 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14161 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	16th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation, SBP-BRiMS 2023
Country/Territory	United States
City	Pittsburgh
Period	9/20/23 → 9/22/23

Keywords

Few-shot Learning
Meta Learning
Network Analysis

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-43129-6_27

Cite this

Mathavan, H., Tan, Z., Mudiam, N., & Liu, H. (2023). Inductive Linear Probing for Few-Shot Node Classification. In R. Thomson, S. Al-khateeb, A. Burger, P. Park, & A. A. Pyke (Eds.), Social, Cultural, and Behavioral Modeling - 16th International Conference, SBP-BRiMS 2023, Proceedings (pp. 274-284). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14161 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-43129-6_27

Inductive Linear Probing for Few-Shot Node Classification. / Mathavan, Hirthik; Tan, Zhen; Mudiam, Nivedh et al.
Social, Cultural, and Behavioral Modeling - 16th International Conference, SBP-BRiMS 2023, Proceedings. ed. / Robert Thomson; Samer Al-khateeb; Annetta Burger; Patrick Park; Aryn A. Pyke. Springer Science and Business Media Deutschland GmbH, 2023. p. 274-284 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14161 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Mathavan, H, Tan, Z, Mudiam, N & Liu, H 2023, Inductive Linear Probing for Few-Shot Node Classification. in R Thomson, S Al-khateeb, A Burger, P Park & A A. Pyke (eds), Social, Cultural, and Behavioral Modeling - 16th International Conference, SBP-BRiMS 2023, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14161 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 274-284, 16th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation, SBP-BRiMS 2023, Pittsburgh, United States, 9/20/23. https://doi.org/10.1007/978-3-031-43129-6_27

Mathavan H, Tan Z, Mudiam N, Liu H. Inductive Linear Probing for Few-Shot Node Classification. In Thomson R, Al-khateeb S, Burger A, Park P, A. Pyke A, editors, Social, Cultural, and Behavioral Modeling - 16th International Conference, SBP-BRiMS 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 274-284. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-43129-6_27

Mathavan, Hirthik ; Tan, Zhen ; Mudiam, Nivedh et al. / Inductive Linear Probing for Few-Shot Node Classification. Social, Cultural, and Behavioral Modeling - 16th International Conference, SBP-BRiMS 2023, Proceedings. editor / Robert Thomson ; Samer Al-khateeb ; Annetta Burger ; Patrick Park ; Aryn A. Pyke. Springer Science and Business Media Deutschland GmbH, 2023. pp. 274-284 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{c78f8496df2f419d8e2960f8d83f3836,

title = "Inductive Linear Probing for Few-Shot Node Classification",

abstract = "Meta-learning has emerged as a powerful training strategy for few-shot node classification, demonstrating its effectiveness in the transductive setting. However, the existing literature predominantly focuses on transductive few-shot node classification, neglecting the widely studied inductive setting in the broader few-shot learning community. This oversight limits our comprehensive understanding of the performance of meta-learning based methods on graph data. In this work, we conduct an empirical study to highlight the limitations of current frameworks in the inductive few-shot node classification setting. Additionally, we propose applying a competitive baseline approach specifically tailored for inductive few-shot node classification tasks. We hope our work can provide a new path forward to better understand how the meta-learning paradigm works in the graph domain.",

keywords = "Few-shot Learning, Meta Learning, Network Analysis",

author = "Hirthik Mathavan and Zhen Tan and Nivedh Mudiam and Huan Liu",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 16th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation, SBP-BRiMS 2023 ; Conference date: 20-09-2023 Through 22-09-2023",

year = "2023",

doi = "10.1007/978-3-031-43129-6_27",

language = "English (US)",

isbn = "9783031431289",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "274--284",

editor = "Robert Thomson and Samer Al-khateeb and Annetta Burger and Patrick Park and {A. Pyke}, Aryn",

booktitle = "Social, Cultural, and Behavioral Modeling - 16th International Conference, SBP-BRiMS 2023, Proceedings",

address = "Germany",

}

TY - GEN

T1 - Inductive Linear Probing for Few-Shot Node Classification

AU - Mathavan, Hirthik

AU - Tan, Zhen

AU - Mudiam, Nivedh

AU - Liu, Huan

PY - 2023

Y1 - 2023

N2 - Meta-learning has emerged as a powerful training strategy for few-shot node classification, demonstrating its effectiveness in the transductive setting. However, the existing literature predominantly focuses on transductive few-shot node classification, neglecting the widely studied inductive setting in the broader few-shot learning community. This oversight limits our comprehensive understanding of the performance of meta-learning based methods on graph data. In this work, we conduct an empirical study to highlight the limitations of current frameworks in the inductive few-shot node classification setting. Additionally, we propose applying a competitive baseline approach specifically tailored for inductive few-shot node classification tasks. We hope our work can provide a new path forward to better understand how the meta-learning paradigm works in the graph domain.

AB - Meta-learning has emerged as a powerful training strategy for few-shot node classification, demonstrating its effectiveness in the transductive setting. However, the existing literature predominantly focuses on transductive few-shot node classification, neglecting the widely studied inductive setting in the broader few-shot learning community. This oversight limits our comprehensive understanding of the performance of meta-learning based methods on graph data. In this work, we conduct an empirical study to highlight the limitations of current frameworks in the inductive few-shot node classification setting. Additionally, we propose applying a competitive baseline approach specifically tailored for inductive few-shot node classification tasks. We hope our work can provide a new path forward to better understand how the meta-learning paradigm works in the graph domain.

KW - Few-shot Learning

KW - Meta Learning

KW - Network Analysis

UR - http://www.scopus.com/inward/record.url?scp=85174449205&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85174449205&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-43129-6_27

DO - 10.1007/978-3-031-43129-6_27

M3 - Conference contribution

AN - SCOPUS:85174449205

SN - 9783031431289

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 274

EP - 284

BT - Social, Cultural, and Behavioral Modeling - 16th International Conference, SBP-BRiMS 2023, Proceedings

A2 - Thomson, Robert

A2 - Al-khateeb, Samer

A2 - Burger, Annetta

A2 - Park, Patrick

A2 - A. Pyke, Aryn

PB - Springer Science and Business Media Deutschland GmbH

T2 - 16th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation, SBP-BRiMS 2023

Y2 - 20 September 2023 through 22 September 2023

ER -