Towards Scalable Lottery Ticket Networks using Genetic Algorithms
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918123826839552 |
|---|---|
| author | Schönberger, Julian Zorn, Maximilian Nüßlein, Jonas Gabor, Thomas Altmann, Philipp |
| author_facet | Schönberger, Julian Zorn, Maximilian Nüßlein, Jonas Gabor, Thomas Altmann, Philipp |
| contents | Building modern deep learning systems that are not just effective but also efficient requires rethinking established paradigms for model training and neural architecture design. Instead of adapting highly overparameterized networks and subsequently applying model compression techniques to reduce resource consumption, a new class of high-performing networks skips the need for expensive parameter updates, while requiring only a fraction of parameters, making them highly scalable. The Strong Lottery Ticket Hypothesis posits that within randomly initialized, sufficiently overparameterized neural networks, there exist subnetworks that can match the accuracy of the trained original model-without any training. This work explores the usage of genetic algorithms for identifying these strong lottery ticket subnetworks. We find that for instances of binary and multi-class classification tasks, our approach achieves better accuracies and sparsity levels than the current state-of-the-art without requiring any gradient information. In addition, we provide justification for the need for appropriate evaluation metrics when scaling to more complex network architectures and learning tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_08877 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Towards Scalable Lottery Ticket Networks using Genetic Algorithms Schönberger, Julian Zorn, Maximilian Nüßlein, Jonas Gabor, Thomas Altmann, Philipp Machine Learning Neural and Evolutionary Computing Building modern deep learning systems that are not just effective but also efficient requires rethinking established paradigms for model training and neural architecture design. Instead of adapting highly overparameterized networks and subsequently applying model compression techniques to reduce resource consumption, a new class of high-performing networks skips the need for expensive parameter updates, while requiring only a fraction of parameters, making them highly scalable. The Strong Lottery Ticket Hypothesis posits that within randomly initialized, sufficiently overparameterized neural networks, there exist subnetworks that can match the accuracy of the trained original model-without any training. This work explores the usage of genetic algorithms for identifying these strong lottery ticket subnetworks. We find that for instances of binary and multi-class classification tasks, our approach achieves better accuracies and sparsity levels than the current state-of-the-art without requiring any gradient information. In addition, we provide justification for the need for appropriate evaluation metrics when scaling to more complex network architectures and learning tasks. |
| title | Towards Scalable Lottery Ticket Networks using Genetic Algorithms |
| topic | Machine Learning Neural and Evolutionary Computing |
| url | https://arxiv.org/abs/2508.08877 |