Summary of Cost-effective Fault Tolerance For Cnns Using Parameter Vulnerability Based Hardening and Pruning, by Mohammad Hasan Ahmadilivani et al.
Cost-Effective Fault Tolerance for CNNs Using Parameter Vulnerability Based Hardening and Pruning
by Mohammad Hasan Ahmadilivani, Seyedhamidreza Mousavi, Jaan Raik, Masoud Daneshtalab, Maksim Jenihhin
First submitted to arxiv on: 17 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel model-level hardening approach for Convolutional Neural Networks (CNNs) is introduced in this paper, which integrates error correction directly into the neural networks. The proposed method, dubbed “error correction layer,” duplicates selective filters/neurons and applies an efficient and robust correction mechanism to ensure fault resilience. While the hardened CNNs demonstrate nearly equivalent performance to Triple Modular Redundancy (TMR)-based correction, they exhibit significantly reduced overhead. To further reduce the computational cost, a cost-effective parameter vulnerability based pruning technique is proposed, which outperforms conventional pruning methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper develops a way to make Convolutional Neural Networks (CNNs) more reliable and fault-tolerant for use in safety-critical applications. The new approach, called “error correction layer,” helps fix mistakes that can occur when the network is processing information. This makes it possible to build more reliable systems using CNNs without needing special hardware or changes to how the network works. |
Keywords
» Artificial intelligence » Pruning
