Papers by Sheng-Wei Chen
Random Label Forests: An Ensemble Method with Label Subsampling For Extreme Multi-Label Problems (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for multi-label learning require large memory space for text classification . recent studies show that multiple labels are needed for e-commerce applications . |
| Approach: | They propose a distributed ensemble method with label subsampling to share large memory space for handling large-scale labels. |
| Outcome: | The proposed method can reduce memory usage while keeping competitive performance over real-world data sets. |