The Oversmoothing Fallacy: A Misguided Narrative in GNN Research
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913877486206976 |
|---|---|
| author | Park, MoonJeong Choi, Sunghyun Heo, Jaeseung Park, Eunhyeok Kim, Dongwoo |
| author_facet | Park, MoonJeong Choi, Sunghyun Heo, Jaeseung Park, Eunhyeok Kim, Dongwoo |
| contents | Oversmoothing has been recognized as a main obstacle to building deep Graph Neural Networks (GNNs), limiting the performance. This position paper argues that the influence of oversmoothing has been overstated and advocates for a further exploration of deep GNN architectures. Given the three core operations of GNNs, aggregation, linear transformation, and non-linear activation, we show that prior studies have mistakenly confused oversmoothing with the vanishing gradient, caused by transformation and activation rather than aggregation. Our finding challenges prior beliefs about oversmoothing being unique to GNNs. Furthermore, we demonstrate that classical solutions such as skip connections and normalization enable the successful stacking of deep GNN layers without performance degradation. Our results clarify misconceptions about oversmoothing and shed new light on the potential of deep GNNs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_04653 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The Oversmoothing Fallacy: A Misguided Narrative in GNN Research Park, MoonJeong Choi, Sunghyun Heo, Jaeseung Park, Eunhyeok Kim, Dongwoo Machine Learning Oversmoothing has been recognized as a main obstacle to building deep Graph Neural Networks (GNNs), limiting the performance. This position paper argues that the influence of oversmoothing has been overstated and advocates for a further exploration of deep GNN architectures. Given the three core operations of GNNs, aggregation, linear transformation, and non-linear activation, we show that prior studies have mistakenly confused oversmoothing with the vanishing gradient, caused by transformation and activation rather than aggregation. Our finding challenges prior beliefs about oversmoothing being unique to GNNs. Furthermore, we demonstrate that classical solutions such as skip connections and normalization enable the successful stacking of deep GNN layers without performance degradation. Our results clarify misconceptions about oversmoothing and shed new light on the potential of deep GNNs. |
| title | The Oversmoothing Fallacy: A Misguided Narrative in GNN Research |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2506.04653 |