Pruning Randomly Initialized Neural Networks With Iterative Randomization
ASecurity**arXiv ID:** 2106.09269 **Authors:** Daiki Chijiwa, Shin'ya Yamaguchi, Yasutoshi Ida, Kenji Umakoshi, Tomohiro Inoue **Published:** 2021-06-17T06:32:57Z **Abstract:** Pruning the weights of randomly initialized neural networks plays an important role in the context of lottery ticket hypothesis. Ramanujan et al. (2020) empirically showed that only pruning the weights can achieve remarkable performance instead of optimizing the weight values. However, to achieve the same level of performance a...
- 3 stars
- 0 votes
- 0 copies
- 4 views
- Added September 11, 2026
Security analysis
100/100npx -y skills add hiyenwong/ai_collection --skill pruning-randomly-initialized-neural-networks-with-iterative-randomization --agent claude-codeAre you the author of Pruning Randomly Initialized Neural Networks With Iterative Randomization?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/hiyenwong-pruning-randomly-initialized-neural-networks-with)# Pruning Randomly Initialized Neural Networks with Iterative Randomization
**arXiv ID:** 2106.09269
**Authors:** Daiki Chijiwa, Shin'ya Yamaguchi, Yasutoshi Ida, Kenji Umakoshi, Tomohiro Inoue
**Published:** 2021-06-17T06:32:57Z
**Abstract:**
Pruning the weights of randomly initialized neural networks plays an important role in the context of lottery ticket hypothesis. Ramanujan et al. (2020) empirically showed that only pruning the weights can achieve remarkable performance instead of optimizing the weight values. However, to achieve the same level of performance as the weight optimization, the pruning approach requires more parameters in the networks before pruning and thus more memory space. To overcome this parameter inefficiency, we introduce a novel framework to prune randomly initialized neural networks with iteratively randomizing weight values (IteRand). Theoretically, we prove an approximation theorem in our framework, which indicates that the randomizing operations are provably effective to reduce the required number of the parameters. We also empirically demonstrate the parameter efficiency in multiple experiments on CIFAR-10 and ImageNet. The code is available at: https://github.com/dchiji-ntt/iterand
## Skill Description
This skill is generated from the arXiv paper: Pruning Randomly Initialized Neural Networks with Iterative Randomization (2106.09269).
## How to Use
[To be filled in by the user or by future automation]
## References
- [arXiv:2106.09269](http://arxiv.org/abs/2106.09269v2)
Attribution
Related skills
View all in tools →Comments
Loading comments…
Newsletter
Get the weekly list of safe new skills, and the ones we pulled.