AdapterGNN: Parameter-Efficient Fine-Tuning Improves Generalization in GNNs
This work addresses the challenge of efficient fine-tuning for GNNs, which is incremental as it adapts PEFT techniques from other fields to improve generalization in graph-based tasks.
The paper tackled the problem of applying parameter-efficient fine-tuning (PEFT) to graph neural networks (GNNs), which was less effective due to differences from transformer-based models, and proposed AdapterGNN, a novel PEFT method that outperformed full fine-tuning by 1.6% and 5.7% in chemistry and biology domains with only 5% and 4% of parameters tuned, while reducing generalization gaps.
Fine-tuning pre-trained models has recently yielded remarkable performance gains in graph neural networks (GNNs). In addition to pre-training techniques, inspired by the latest work in the natural language fields, more recent work has shifted towards applying effective fine-tuning approaches, such as parameter-efficient fine-tuning (PEFT). However, given the substantial differences between GNNs and transformer-based models, applying such approaches directly to GNNs proved to be less effective. In this paper, we present a comprehensive comparison of PEFT techniques for GNNs and propose a novel PEFT method specifically designed for GNNs, called AdapterGNN. AdapterGNN preserves the knowledge of the large pre-trained model and leverages highly expressive adapters for GNNs, which can adapt to downstream tasks effectively with only a few parameters, while also improving the model's generalization ability. Extensive experiments show that AdapterGNN achieves higher performance than other PEFT methods and is the only one consistently surpassing full fine-tuning (outperforming it by 1.6% and 5.7% in the chemistry and biology domains respectively, with only 5% and 4% of its parameters tuned) with lower generalization gaps. Moreover, we empirically show that a larger GNN model can have a worse generalization ability, which differs from the trend observed in large transformer-based models. Building upon this, we provide a theoretical justification for PEFT can improve generalization of GNNs by applying generalization bounds. Our code is available at https://github.com/Lucius-lsr/AdapterGNN.