Beyond Preserved Accuracy: Evaluating Loyalty and Robustness of BERT Compression
arXiv:2109.03228
Abstract
Recent studies on compression of pretrained language models (e.g., BERT) usually use preserved accuracy as the metric for evaluation. In this paper, we propose two new metrics, label loyalty and probability loyalty that measure how closely a compressed model (i.e., student) mimics the original model (i.e., teacher). We also explore the effect of compression with regard to robustness under adversarial attacks. We benchmark quantization, pruning, knowledge distillation and progressive module replacing with loyalty and robustness. By combining multiple compression techniques, we provide a practical strategy to achieve better accuracy, loyalty and robustness.
arXiv commentAccepted to EMNLP 2021 (main conference)
References in corpus (6)
- Distilling the Knowledge in a Neural Network
- A Survey of Model Compression and Acceleration for Deep Neural Networks
- Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism
- MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers
- Challenges in Deploying Machine Learning: a Survey of Case Studies
- Characterising Bias in Compressed Models