Improving Language Understanding by Generative Pre-Training

Natural language understanding comprises a wide range of diverse tasks such as textual entailment, question answering, semantic similarity assessment, and document classification. Although large unlabeled text corpora are abundant, labeled data for learning these specific tasks is scarce, making it challenging for discriminatively trained models to perform adequately. We demonstrate that large gains on these tasks can be realized by generative pre-training of a language model on a diverse corpus of unlabeled text, followed by discriminative fine-tuning on each specific task. In contrast to previous approaches, we make use of task-aware input transformations during fine-tuning to achieve effective transfer while requiring minimal changes to the model architecture. We demonstrate the effectiveness of our approach on a wide range of benchmarks for natural language understanding. Our general task-agnostic model outperforms discriminatively trained models that use architectures specifically crafted for each task, significantly improving upon the state of the art in 9 out of the 12 tasks studied. For instance, we achieve absolute improvements of 8.9% on commonsense reasoning (Stories Cloze Test), 5.7% on question answering (RACE), and 1.5% on textual entailment (MultiNLI).


Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Natural Language Inference MultiNLI Finetuned Transformer LM Matched 82.1 # 38
Mismatched 81.4 # 29
Question Answering RACE Finetuned Transformer LM RACE-m 62.9 # 4
RACE-h 57.4 # 3
RACE 59.0 # 4
Natural Language Inference SciTail Finetuned Transformer LM Accuracy 88.3 # 3
Natural Language Inference SNLI Fine-Tuned LM-Pretrained Transformer % Test Accuracy 89.9 # 13
% Train Accuracy 96.6 # 5
Parameters 85m # 4
Question Answering StoryCloze Finetuned Transformer LM Accuracy 86.5 # 8