Skip to yearly menu bar Skip to main content


De-confounded Data-free Knowledge Distillation for Handling Distribution Shifts

Yuzheng Wang · Dingkang Yang · Zhaoyu Chen · Yang Liu · Siao Liu · Wenqiang Zhang · Lihua Zhang · Lizhe Qi

Arch 4A-E Poster #289
[ ]
Thu 20 Jun 10:30 a.m. PDT — noon PDT


Data-Free Knowledge Distillation (DFKD) is a promising task to train high-performance small models to enhance actual deployment without relying on the original training data.Existing methods commonly avoid relying on private data by utilizing synthetic or sampled data.However, a long-overlooked issue is that the severe distribution shifts between their substitution and original data, which manifests as huge differences in the quality of images and class proportions.The harmful shifts are essentially the confounder that significantly causes performance bottlenecks.To tackle the issue, this paper proposes a novel perspective with causal inference to disentangle the student models from the impact of such shifts.By designing a customized causal graph, we first reveal the causalities among the variables in the DFKD task.Subsequently, we propose a Knowledge Distillation Causal Intervention (KDCI) framework based on the backdoor adjustment to de-confound the confounder.KDCI can be flexibly combined with most existing state-of-the-art baselines. Experiments in combination with six representative DFKD methods demonstrate the effectiveness of our KDCI, which can obviously help existing methods under almost all settings, e.g., improving the baseline by up to 15.54\% accuracy on the CIFAR-100 dataset.

Live content is unavailable. Log in and register to view live content