Papers › Alleviating Exposure Bias via Multi-level Contrastive Learning and Deviation...

Alleviating Exposure Bias via Multi-level Contrastive Learning and Deviation Simulation in Abstractive Summarization

3 May 2023ACL 2023 5archive 2025-07-28

Jiawen Xie, Qi Su, Shaoting Zhang, and Xiaofan Zhang

Most Transformer based abstractive summarization systems have a severe mismatch between training and inference, i.e., exposure bias. From diverse perspectives, we introduce a simple multi-level contrastive learning framework for abstractive summarization (SimMCS) and a tailored sparse decoder self-attention pattern (SDSA) to bridge the gap between training and inference to improve model performance. Compared with previous contrastive objectives focusing only on the relative order of probability mass assigned to non-gold summaries, SimMCS additionally takes their absolute positions into account, which guarantees that the relatively high-quality (positive) summaries among them could be properly assigned high probability mass, and further enhances the capability of discriminating summary quality beyond exploiting potential artifacts of specific metrics. SDSA simulates the possible inference scenarios of deviation in the training phase to get closer to the ideal paradigm. Our approaches outperform the previous state-of-the-art results on two summarization datasets while just adding fairly low overhead. Further empirical analysis shows our model preserves the advantages of prior contrastive methods and possesses strong few-shot learning ability.

Paper PDF Code

Code

xjw-nlp/SimMCS pytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Abstractive Text Summarization Contrastive Learning Decoder Few-Shot Learning

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Absolute Position Encodings Adam Attention BPE Contrastive Learning Dense Connections Dropout Label Smoothing Layer Normalization Linear Layer Multi-Head Attention Position-Wise Feed-Forward Layer Residual Connection Softmax Transformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections