2 |
Optimizing Deeper Transformers on Small Datasets ...
|
|
The Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing 2021; Cao, Yanshuai; Cheung, Jackie Chi Kit; Huang , Chenyang; Kumar, Dhruv; Prince, Simon; Tang, Keyi; Xu, Peng; Yang, Wei; Zi, Wenjie. - : Underline Science Inc., 2021
|
|
Abstract:
Read paper: https://www.aclanthology.org/2021.acl-long.163 Abstract: It is a common belief that training deep transformers from scratch requires large datasets. Consequently, for small datasets, people usually use shallow and simple additional layers on top of pre-trained models during fine-tuning. This work shows that this does not always need to be the case: with proper initialization and optimization, the benefits of very deep transformers can carry over to challenging tasks with small datasets, including Text-to-SQL semantic parsing and logical reading comprehension. In particular, we successfully train 48 layers of transformers, comprising 24 fine-tuned layers from pre-trained RoBERTa and 24 relation-aware layers trained from scratch. With fewer training steps and no task-specific pre-training, we obtain the state of the art performance on the challenging cross-domain Text-to-SQL parsing benchmark Spider. We achieve this by deriving a novel Data dependent Transformer Fixed-update initialization scheme ...
|
|
Keyword:
Computational Linguistics; Condensed Matter Physics; Deep Learning; Electromagnetism; FOS Physical sciences; Information and Knowledge Engineering; Neural Network; Semantics
|
|
URL: https://dx.doi.org/10.48448/ehsy-3055 https://underline.io/lecture/25482-optimizing-deeper-transformers-on-small-datasets
|
|
BASE
|
|
Hide details
|
|
3 |
Textual Time Travel: A Temporally Informed Approach to Theory of Mind ...
|
|
|
|
BASE
|
|
Show details
|
|
4 |
Modeling Event Plausibility with Consistent Conceptual Abstraction ...
|
|
|
|
BASE
|
|
Show details
|
|
7 |
On the Systematicity of Probing Contextualized Word Representations: The Case of Hypernymy in BERT ...
|
|
|
|
BASE
|
|
Show details
|
|
8 |
Learning Efficient Task-Specific Meta-Embeddings with Word Prisms ...
|
|
|
|
BASE
|
|
Show details
|
|
9 |
Resolving Event Coreference with Supervised Representation Learning and Clustering-Oriented Regularization ...
|
|
|
|
BASE
|
|
Show details
|
|
10 |
Leveraging Lexical Resources for Learning Entity Embeddings in Multi-Relational Data ...
|
|
|
|
BASE
|
|
Show details
|
|
11 |
Distributional Semantics for Robust Automatic Summarization
|
|
|
|
BASE
|
|
Show details
|
|
|
|