MétaCan
Menu
Back to cohort
Record W7134973199 · doi:10.1109/icdmw69685.2025.00120

Smooth Transitions in Graph Self-Supervision: Mitigating Feature Twist Across Abstraction Levels

2025· article· W7134973199 on OpenAlex

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

affAt least one author lists a Canadian institution in the pinned OpenAlex snapshot.

Bibliographic record

Venuenot available
Typearticle
Language
FieldComputer Science
TopicAdvanced Graph Neural Networks
Canadian institutionsUniversité du Québec à Montréal
Fundersnot available
KeywordsFeature (linguistics)TwistAbstractionGraphContext (archaeology)

Abstract

fetched live from OpenAlex

Self-supervised learning has emerged as a powerful paradigm for graph representation learning, enabling models to leverage structural and feature information without manual labels. Despite its success, the geometric challenges posed by transitions across abstraction levels, including node, proximity, cluster, and graph granularities, remain underexplored. This paper introduces a novel self-supervised contrastive learning framework and investigates the evolution of intrinsic dimensions during these transitions. Leveraging a pretraining and finetuning protocol, we highlight a critical challenge termed Feature Twist, characterized by abrupt geometric distortions in the latent space, leading to performance degradation in downstream tasks. To address this, we propose a robust filtering mechanism that smooths transitions by mitigating these distortions. Experimental evaluations on benchmark datasets demonstrate the competitive advantages of our approach across common downstream tasks, including node classification, node clustering, and link prediction. Our findings underscore the importance of geometric considerations in self-supervised graph representation learning.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.001
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMeta-epidemiology (narrow)
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Other design · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.804
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0010.000
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0010.000
Bibliometrics0.0000.006
Science and technology studies0.0010.000
Scholarly communication0.0010.002
Open science0.0010.000
Research integrity0.0010.002
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.017
GPT teacher head0.297
Teacher spread0.280 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Quick stats

Citations0
Published2025
Admission routes1
Has abstractyes

Explore more

Same topicAdvanced Graph Neural NetworksFrench-language works237,207