Summary of A Federated Learning Approach to Privacy Preserving Offensive Language Identification, by Marcos Zampieri et al.
A Federated Learning Approach to Privacy Preserving Offensive Language Identification
by Marcos Zampieri, Damith Premasiri, Tharindu Ranasinghe
First submitted to arxiv on: 17 Apr 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Machine Learning (cs.LG)
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 proposes a decentralized architecture for identifying offensive language online, focusing on privacy preservation. The authors introduce Federated Learning (FL) to train multiple models locally without sharing user data, addressing a crucial concern in social media. By applying FL in the context of offensive language identification, the model fusion approach outperforms baselines in four English benchmark datasets (AHSD, HASOC, HateXplain, OLID). Initial cross-lingual experiments are also presented in English and Spanish. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper solves a big problem on social media: stopping mean speech online while keeping people’s privacy safe. Right now, models that detect mean language are trained using lots of data stored in central servers. This isn’t private because most social media data comes from users. The authors suggest a new way to train these models without sharing user data. They call it Federated Learning (FL). FL lets multiple models be trained locally without sharing data, keeping privacy safe. The authors tested their idea using four big language datasets and showed that it works better than usual methods. |
Keywords
» Artificial intelligence » Federated learning
