Where Did Your Model Learn That? Label-free Influence for Self-supervised Learning
Guardat en:
| Publicat a: | arXiv.org (Dec 22, 2024), p. n/a |
|---|---|
| Autor principal: | |
| Altres autors: | , , , |
| Publicat: |
Cornell University Library, arXiv.org
|
| Matèries: | |
| Accés en línia: | Citation/Abstract Full text outside of ProQuest |
| Etiquetes: |
Sense etiquetes, Sigues el primer a etiquetar aquest registre!
|
MARC
| LEADER | 00000nab a2200000uu 4500 | ||
|---|---|---|---|
| 001 | 3148949001 | ||
| 003 | UK-CbPIL | ||
| 022 | |a 2331-8422 | ||
| 035 | |a 3148949001 | ||
| 045 | 0 | |b d20241222 | |
| 100 | 1 | |a Harilal, Nidhin | |
| 245 | 1 | |a Where Did Your Model Learn That? Label-free Influence for Self-supervised Learning | |
| 260 | |b Cornell University Library, arXiv.org |c Dec 22, 2024 | ||
| 513 | |a Working Paper | ||
| 520 | 3 | |a Self-supervised learning (SSL) has revolutionized learning from large-scale unlabeled datasets, yet the intrinsic relationship between pretraining data and the learned representations remains poorly understood. Traditional supervised learning benefits from gradient-based data attribution tools like influence functions that measure the contribution of an individual data point to model predictions. However, existing definitions of influence rely on labels, making them unsuitable for SSL settings. We address this gap by introducing Influence-SSL, a novel and label-free approach for defining influence functions tailored to SSL. Our method harnesses the stability of learned representations against data augmentations to identify training examples that help explain model predictions. We provide both theoretical foundations and empirical evidence to show the utility of Influence-SSL in analyzing pre-trained SSL models. Our analysis reveals notable differences in how SSL models respond to influential data compared to supervised models. Finally, we validate the effectiveness of Influence-SSL through applications in duplicate detection, outlier identification and fairness analysis. Code is available at: \url{https://github.com/cryptonymous9/Influence-SSL}. | |
| 653 | |a Outliers (statistics) | ||
| 653 | |a Data analysis | ||
| 653 | |a Data augmentation | ||
| 653 | |a Self-supervised learning | ||
| 653 | |a Labels | ||
| 653 | |a Influence functions | ||
| 653 | |a Harnesses | ||
| 653 | |a Representations | ||
| 653 | |a Data points | ||
| 653 | |a Stability augmentation | ||
| 700 | 1 | |a Rege, Amit Kiran | |
| 700 | 1 | |a Reza Akbarian Bafghi | |
| 700 | 1 | |a Raissi, Maziar | |
| 700 | 1 | |a Monteleoni, Claire | |
| 773 | 0 | |t arXiv.org |g (Dec 22, 2024), p. n/a | |
| 786 | 0 | |d ProQuest |t Engineering Database | |
| 856 | 4 | 1 | |3 Citation/Abstract |u https://www.proquest.com/docview/3148949001/abstract/embedded/ZKJTFFSVAI7CB62C?source=fedsrch |
| 856 | 4 | 0 | |3 Full text outside of ProQuest |u http://arxiv.org/abs/2412.17170 |