Don't Pay Attention, PLANT It: Pretraining Attention via Learning-to-Rank
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915693537001472 |
|---|---|
| author | Roy, Debjyoti Saha Wallace, Byron C. Aslam, Javed A. |
| author_facet | Roy, Debjyoti Saha Wallace, Byron C. Aslam, Javed A. |
| contents | State-of-the-art Extreme Multi-Label Text Classification models rely on multi-label attention to focus on key tokens in input text, but learning good attention weights is challenging. We introduce PLANT - Pretrained and Leveraged Attention - a plug-and-play strategy for initializing attention. PLANT works by planting label-specific attention using a pretrained Learning-to-Rank model guided by mutual information gain. This architecture-agnostic approach integrates seamlessly with large language model backbones such as Mistral-7B, LLaMA3-8B, DeepSeek-V3, and Phi-3. PLANT outperforms state-of-the-art methods across tasks including ICD coding, legal topic classification, and content recommendation. Gains are especially pronounced in few-shot settings, with substantial improvements on rare labels. Ablation studies confirm that attention initialization is a key driver of these gains. For code and trained models, see https://github.com/debjyotiSRoy/xcube/tree/plant |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_23066 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Don't Pay Attention, PLANT It: Pretraining Attention via Learning-to-Rank Roy, Debjyoti Saha Wallace, Byron C. Aslam, Javed A. Computation and Language Machine Learning State-of-the-art Extreme Multi-Label Text Classification models rely on multi-label attention to focus on key tokens in input text, but learning good attention weights is challenging. We introduce PLANT - Pretrained and Leveraged Attention - a plug-and-play strategy for initializing attention. PLANT works by planting label-specific attention using a pretrained Learning-to-Rank model guided by mutual information gain. This architecture-agnostic approach integrates seamlessly with large language model backbones such as Mistral-7B, LLaMA3-8B, DeepSeek-V3, and Phi-3. PLANT outperforms state-of-the-art methods across tasks including ICD coding, legal topic classification, and content recommendation. Gains are especially pronounced in few-shot settings, with substantial improvements on rare labels. Ablation studies confirm that attention initialization is a key driver of these gains. For code and trained models, see https://github.com/debjyotiSRoy/xcube/tree/plant |
| title | Don't Pay Attention, PLANT It: Pretraining Attention via Learning-to-Rank |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2410.23066 |