Self-Supervised Representation Learning on Neural Network Weights for Model Characteristic Prediction
Konstantin Schürholt, Dimche Kostadinov, Damian Borth
Abstract
Self-Supervised Learning (SSL) has been shown to learn useful and information-preserving representations. Neural Networks (NNs) are widely applied, yet their weight space is still not fully understood. Therefore, we propose to use SSL to learn hyper-representations of the weights of populations of NNs. To that end, we introduce domain specific data augmentations and an adapted attention architecture. Our empirical evaluation demonstrates that self-supervised representation learning in this domain is able to recover diverse NN model characteristics. Further, we show that the proposed learned representations outperform prior work for predicting hyper-parameters, test accuracy, and generalization gap as well as transfer to out-of-distribution settings.
BibTeX
@inproceedings{
sch{\"u}rholt2021selfsupervised,
title={Self-Supervised Representation Learning on Neural Network Weights for Model Characteristic Prediction},
author={Konstantin Sch{\"u}rholt and Dimche Kostadinov and Damian Borth},
booktitle={Advances in Neural Information Processing Systems},
editor={A. Beygelzimer and Y. Dauphin and P. Liang and J. Wortman Vaughan},
year={2021},
url={https://openreview.net/forum?id=F1D8buayXQT}
}