论文标题
实体抽象是否有助于生成变压器的推理？
Does Entity Abstraction Help Generative Transformers Reason?
论文作者
论文摘要
我们研究了将实体类型抽象纳入预训练的变压器中的实用性，并在需要不同形式的逻辑推理的四个NLP任务上测试这些方法：（1）与基于文本的关系推理（CLUTRR），（2）绑架性推理（2）绑架词（2）证明作者（证明器），（3）多跳问题答案（3）Multi-Hop Question（HOTPOTQA）和（4）copation（4）对话（4）对话（4）对话（4）对话（4）对话（4）conditation（4）对话（4）对话（4）对话（4）对话（4）对话（4）对话（4）对话（4）对话（4）对话（4）conection（4）对话（4）对话（4）conditation（4）condication（4），我们提出并经验探索添加这种抽象的三种方法：（i）作为其他输入嵌入，（ii）作为编码的单独序列，（iii）作为模型的辅助预测任务。总体而言，我们的分析表明，具有抽象实体知识的模型比没有它的模型表现更好。最佳抽象意识模型的总体准确度为88.8％和91.8％，而基线模型分别达到62.9％和89.8％，分别在Clutrr和证明作者方面达到了89.8％。但是，对于HOTPOTQA和COQA，我们发现F1得分平均仅提高0.5％。我们的结果表明，明确抽象的好处在正式定义的逻辑推理设置需要许多推理之上具有重要意义，但要指出这样的观点是，它对NLP任务的好处不太有益。
We study the utility of incorporating entity type abstractions into pre-trained Transformers and test these methods on four NLP tasks requiring different forms of logical reasoning: (1) compositional language understanding with text-based relational reasoning (CLUTRR), (2) abductive reasoning (ProofWriter), (3) multi-hop question answering (HotpotQA), and (4) conversational question answering (CoQA). We propose and empirically explore three ways to add such abstraction: (i) as additional input embeddings, (ii) as a separate sequence to encode, and (iii) as an auxiliary prediction task for the model. Overall, our analysis demonstrates that models with abstract entity knowledge performs better than without it. The best abstraction aware models achieved an overall accuracy of 88.8% and 91.8% compared to the baseline model achieving 62.9% and 89.8% on CLUTRR and ProofWriter respectively. However, for HotpotQA and CoQA, we find that F1 scores improve by only 0.5% on average. Our results suggest that the benefit of explicit abstraction is significant in formally defined logical reasoning settings requiring many reasoning hops, but point to the notion that it is less beneficial for NLP tasks having less formal logical structure.
