Noisy Self-Knowledge Distillation for Text Summarization
arXiv:2009.07032
Abstract
In this paper we apply self-knowledge distillation to text summarization which we argue can alleviate problems with maximum-likelihood training on single reference and noisy datasets. Instead of relying on one-hot annotation labels, our student summarization model is trained with guidance from a teacher which generates smoothed labels to help regularize training. Furthermore, to better model uncertainty during training, we introduce multiple noise signals for both teacher and student models. We demonstrate experimentally on three benchmarks that our framework boosts the performance of both pretrained and non-pretrained summarizers achieving state-of-the-art results.
arXiv commentupdate proceeding version
References in corpus (17)
- Distilling the Knowledge in a Neural Network
- Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
- Teaching Machines to Read and Comprehend
- Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer
- Do Deep Nets Really Need to be Deep?
- A Deep Reinforced Model for Abstractive Summarization
- MASS: Masked Sequence to Sequence Pre-training for Language Generation
- Self-training with Noisy Student improves ImageNet classification
- UniLMv2: Pseudo-Masked Language Models for Unified Language Model Pre-Training
- Revisiting Self-Training for Neural Sequence Generation
- Towards Understanding Knowledge Distillation
- Multilingual Neural Machine Translation with Knowledge Distillation
- Do Deep Convolutional Nets Really Need to be Deep and Convolutional?
- Self-Distillation Amplifies Regularization in Hilbert Space
- Structure-Infused Copy Mechanisms for Abstractive Summarization
- Word Embedding Perturbation for Sentence Classification
- Self-Knowledge Distillation in Natural Language Processing
Cited by in corpus (5)
- Pre-trained Summarization Distillation
- Understanding and Improving Lexical Choice in Non-Autoregressive Translation
- DialogSum: A Real-Life Scenario Dialogue Summarization Dataset
- Multilingual AMR Parsing with Noisy Knowledge Distillation
- Combining Different V1 Brain Model Variants to Improve Robustness to Image Corruptions in CNNs