# Better Pre-Training by Reducing Representation Confusion

arXiv: https://arxiv.org/abs/2210.04246 · HTML: https://mlsanity.com/paper/2210.04246/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 32 references that are in the map.
- First posted on arXiv 9 Oct 2022.

## Builds on (most cited first, 25 of 25)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://mlsanity.com/paper/1906.08237/) (2019)
- [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://mlsanity.com/paper/1909.11942/) (2019)
- [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://mlsanity.com/paper/2006.03654/) (2020)
- [Decoupled Weight Decay Regularization](https://mlsanity.com/paper/1711.05101/) (2017)
- [Self-Attention with Relative Position Representations](https://mlsanity.com/paper/1803.02155/) (2018)
- [Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books](https://mlsanity.com/paper/1506.06724/) (2015)
- [What Does BERT Look At? An Analysis of BERT's Attention](https://mlsanity.com/paper/1906.04341/) (2019)
- [Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned](https://mlsanity.com/paper/1905.09418/) (2019)
- [SpanBERT: Improving Pre-training by Representing and Predicting Spans](https://mlsanity.com/paper/1907.10529/) (2019)
- [Are Sixteen Heads Really Better than One?](https://mlsanity.com/paper/1905.10650/) (2019)
- [Revealing the Dark Secrets of BERT](https://mlsanity.com/paper/1908.08593/) (2019)
- [All-but-the-Top: Simple and Effective Postprocessing for Word Representations](https://mlsanity.com/paper/1702.01417/) (2017)
- [Rethinking Positional Encoding in Language Pre-training](https://mlsanity.com/paper/2006.15595/) (2020)
- [Adaptively Sparse Transformers](https://mlsanity.com/paper/1909.00015/) (2019)
- [DiSAN: Directional Self-Attention Network for RNN/CNN-Free Language Understanding](https://mlsanity.com/paper/1709.04696/) (2017)
- [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://mlsanity.com/paper/1909.00204/) (2019)
- [Parameter Sharing Methods for Multilingual Self-Attentional Translation Models](https://mlsanity.com/paper/1809.00252/) (2018)
- [Pay Better Attention to Attention: Head Selection in Multilingual and Multi-Domain Sequence Modeling](https://mlsanity.com/paper/2106.10840/) (2021)

## Abstract (arXiv)

In this work, we revisit the Transformer-based pre-trained language models and identify two different types of information confusion in position encoding and model representations, respectively. Firstly, we show that in the relative position encoding, the joint modeling about relative distances and directions brings confusion between two heterogeneous information. It may make the model unable to capture the associative semantics of the same distance and the opposite directions, which in turn affects the performance of downstream tasks. Secondly, we notice the BERT with Mask Language Modeling (MLM) pre-training objective outputs similar token representations (last hidden states of different tokens) and head representations (attention weights of different heads), which may make the diversity of information expressed by different tokens and heads limited. Motivated by the above investigation, we propose two novel techniques to improve pre-trained language models: Decoupled Directional Relative Position (DDRP) encoding and MTH pre-training objective. DDRP decouples the relative distance features and the directional features in classical relative position encoding. MTH applies two novel auxiliary regularizers besides MLM to enlarge the dissimilarities between (a) last hidden states of different tokens, and (b) attention weights of different heads. These designs allow the model to capture different categories of information more clearly, as a way to alleviate information confusion in representation learning for better optimization. Extensive experiments and ablation studies on GLUE benchmark demonstrate the effectiveness of our proposed methods.