Yang Liu, Chengjie Sun, Lei Lin, Xiaolong Wang
In this paper, we proposed a sentence encoding-based model for recognizing text entailment. In our approach, the encoding of sentence is a two-stage process. Firstly, average pooling was used over word-level bidirectional LSTM (biLSTM) to generate a first-stage sentence representation. Secondly, attention mechanism was employed to replace average pooling on the same sentence for better representations. Instead of using target sentence to attend words in source sentence, we utilized the sentence's first-stage representation to attend words appeared in itself, which is called "Inner-Attention" in our paper . Experiments conducted on Stanford Natural Language Inference (SNLI) Corpus has proved the effectiveness of "Inner-Attention" mechanism. With less number of parameters, our model outperformed the existing best sentence encoding-based approach by a large margin.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Natural Language Inference | SNLI | % Test Accuracy | 85 | 600D (300+300) BiLSTM encoders with intra-attention and symbolic preproc. |
| Natural Language Inference | SNLI | % Train Accuracy | 85.9 | 600D (300+300) BiLSTM encoders with intra-attention and symbolic preproc. |
| Natural Language Inference | SNLI | % Test Accuracy | 84.2 | 600D (300+300) BiLSTM encoders with intra-attention |
| Natural Language Inference | SNLI | % Train Accuracy | 84.5 | 600D (300+300) BiLSTM encoders with intra-attention |
| Natural Language Inference | SNLI | % Test Accuracy | 83.3 | 600D (300+300) BiLSTM encoders |
| Natural Language Inference | SNLI | % Train Accuracy | 86.4 | 600D (300+300) BiLSTM encoders |