Does Entity Abstraction Help Generative Transformers Reason?
This work addresses the problem of enhancing logical reasoning in NLP models for researchers and practitioners, but it is incremental as it explores variations of existing methods without introducing a fundamentally new approach.
The study investigated whether incorporating entity type abstractions into pre-trained Transformers improves performance on NLP tasks requiring logical reasoning, finding that it significantly boosts accuracy on tasks like CLUTRR (from 62.9% to 88.8%) and ProofWriter (from 89.8% to 91.8%), but yields minimal gains (0.5% F1 improvement) on less formally structured tasks like HotpotQA and CoQA.
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.