Summary of Data-algorithm-architecture Co-optimization For Fair Neural Networks on Skin Lesion Dataset, by Yi Sheng et al.
Data-Algorithm-Architecture Co-Optimization for Fair Neural Networks on Skin Lesion Dataset
by Yi Sheng, Junhuan Yang, Jinyang Li, James Alaina, Xiaowei Xu, Yiyu Shi, Jingtong Hu, Weiwen Jiang, Lei Yang
First submitted to arxiv on: 18 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 This paper highlights the critical concern of fairness in Artificial Intelligence (AI), particularly in medical AI where datasets often reflect inherent biases due to social factors. The authors argue that traditional approaches to mitigating these biases, such as data augmentation and fairness-aware training algorithms, are insufficient. Instead, they propose a holistic approach that considers data, algorithms, and architecture. Utilizing Neural Architecture Search (NAS) technology, specifically Automated ML (AutoML), the paper introduces BiaslessNAS, a novel framework designed to achieve fair outcomes in analyzing skin lesion datasets. The authors demonstrate that BiaslessNAS can identify neural networks that are both more accurate and significantly fairer than traditional NAS methods, with a 2.55% increase in accuracy and a 65.50% improvement in fairness. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making sure artificial intelligence (AI) is fair to everyone. Right now, AI can be biased because the data it’s trained on doesn’t represent all groups of people equally. The authors think this is a big problem, especially when AI is used in medicine. They’re proposing a new way to make AI fair by considering not just the data and algorithms, but also how the AI is designed. This approach uses something called Neural Architecture Search (NAS) to find the best design for an AI system that is both accurate and fair. |
Keywords
» Artificial intelligence » Data augmentation
