IMO: Greedy Layer-Wise Sparse Representation Learning for Out-of-Distribution Text Classification with Pre-trained Models
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911847661174784 |
|---|---|
| author | Feng, Tao Qu, Lizhen Li, Zhuang Zhan, Haolan Hua, Yuncheng Haffari, Gholamreza |
| author_facet | Feng, Tao Qu, Lizhen Li, Zhuang Zhan, Haolan Hua, Yuncheng Haffari, Gholamreza |
| contents | Machine learning models have made incredible progress, but they still struggle when applied to examples from unseen domains. This study focuses on a specific problem of domain generalization, where a model is trained on one source domain and tested on multiple target domains that are unseen during training. We propose IMO: Invariant features Masks for Out-of-Distribution text classification, to achieve OOD generalization by learning invariant features. During training, IMO would learn sparse mask layers to remove irrelevant features for prediction, where the remaining features keep invariant. Additionally, IMO has an attention module at the token level to focus on tokens that are useful for prediction. Our comprehensive experiments show that IMO substantially outperforms strong baselines in terms of various evaluation metrics and settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_13504 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | IMO: Greedy Layer-Wise Sparse Representation Learning for Out-of-Distribution Text Classification with Pre-trained Models Feng, Tao Qu, Lizhen Li, Zhuang Zhan, Haolan Hua, Yuncheng Haffari, Gholamreza Computation and Language Machine learning models have made incredible progress, but they still struggle when applied to examples from unseen domains. This study focuses on a specific problem of domain generalization, where a model is trained on one source domain and tested on multiple target domains that are unseen during training. We propose IMO: Invariant features Masks for Out-of-Distribution text classification, to achieve OOD generalization by learning invariant features. During training, IMO would learn sparse mask layers to remove irrelevant features for prediction, where the remaining features keep invariant. Additionally, IMO has an attention module at the token level to focus on tokens that are useful for prediction. Our comprehensive experiments show that IMO substantially outperforms strong baselines in terms of various evaluation metrics and settings. |
| title | IMO: Greedy Layer-Wise Sparse Representation Learning for Out-of-Distribution Text Classification with Pre-trained Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2404.13504 |