컴퓨터융합학부

축하 드립니다.

1. 수상 정보: 한국정보과학회 2023 한국소프트웨어종합학술대회 2023 우수발표논문상

2. 수상자(저자): Penjan Antonio Eng Lim, 강성원, 박정희(지도교수)

3. 논문 제목: Collaborative Ensemble Construction for a Federated Random Forest

                    under Non-IID Data Distributions

4. 논문 요약: 

Random forests are fundamental in machine learning for their robustness and versatility. However, their centralized nature often poses challenges when dealing with distributed, decentralized, and private data sources. Federated learning (FL) has emerged as a paradigm allowing distributed learning without compromising privacy. Nevertheless, porting ensemble methods like random forests into the federated environment introduces unique challenges, especially in handling non-IID data distributions. This paper presents a federated random forest approach that employs a novel ensemble construction method. Rather than growing trees independently in each client, our approach ensures each decision tree in the ensemble is collectively constructed across clients, with each client accumulating class frequency distributions in the leaf nodes based on its local data. This strategy allows the ensemble to better reflect the data's heterogeneity across different clients, enhancing its performance on non-IID data, as our experimental results confirm.