提出Cognac方法，高效修正图神经网络中错误数据的影响
A Cognac Shot To Forget Bad Memories: Corrective Unlearning for Graph Neural Networks
- 基于消息传递机制设计可逆修正策略，仅需5%误标数据即可修复
- 在真实数据集上恢复近90%原模型性能，优于重训练且快8倍
- 适合处理图数据中恶意或错误节点的后期修正场景
图神经网络广泛应用于图数据的机器学习任务。由于图数据不满足独立同分布假设，恶意篡改或错误数据会通过消息传递传播至其他节点，损害模型性能。为帮助模型开发者在训练后消除被操纵实体的负面影响，本文研究了近期提出的纠正性遗忘问题。我们发现现有图遗忘方法即使已知全部篡改集合，仍无法有效消除影响。为此提出新方法Cognac，可在仅识别出5%篡改数据时，恢复接近使用完整修正训练数据的强基线模型性能。其效果不仅优于从头重新训练（不含删除集），且效率提升8倍。我们希望该工作能助力实际应用中图数据质量问题的后期治理。代码已开源。
原文摘要 · Abstract (English)
Graph Neural Networks (GNNs) are increasingly being used for a variety of ML applications on graph data. Because graph data does not follow the independently and identically distributed (i.i.d.) assumption, adversarial manipulations or incorrect data can propagate to other data points through message passing, which deteriorates the model's performance. To allow model developers to remove the adverse effects of manipulated entities from a trained GNN, we study the recently formulated problem of Corrective Unlearning. We find that current graph unlearning methods fail to unlearn the effect of manipulations even when the whole manipulated set is known. We introduce a new graph unlearning method, Cognac, which can unlearn the effect of the manipulation set even when only 5% of it is identified. It recovers most of the performance of a strong oracle with fully corrected training data, even beating retraining from scratch without the deletion set while being 8x more efficient. We hope our work assists GNN developers in mitigating harmful effects caused by issues in real-world data, post-training. Our code is publicly available at https://github.com/cognac-gnn-unlearning/corrective-unlearning-for-gnns
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。