We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Human Parity on CommonsenseQA: Augmenting Self-Attention with External Attention

Abstract: Most of today's AI systems focus on using self-attention mechanisms and transformer architectures on large amounts of diverse data to achieve impressive performance gains. In this paper, we propose to augment the transformer architecture with an external attention mechanism to bring external knowledge and context to bear. By integrating external information into the prediction process, we hope to reduce the need for ever-larger models and increase the democratization of AI systems. We find that the proposed external attention mechanism can significantly improve the performance of existing AI systems, allowing practitioners to easily customize foundation AI models to many diverse downstream applications. In particular, we focus on the task of Commonsense Reasoning, demonstrating that the proposed external attention mechanism can augment existing transformer models and significantly improve the model's reasoning capabilities. The proposed system, Knowledgeable External Attention for commonsense Reasoning (KEAR), reaches human parity on the open CommonsenseQA research benchmark with an accuracy of 89.4\% in comparison to the human accuracy of 88.9\%.
Comments: 8 pages, 1 figure, 7 tables
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Journal reference: IJCAI 2022 (Long oral)
Cite as: arXiv:2112.03254 [cs.CL]
  (or arXiv:2112.03254v3 [cs.CL] for this version)

Submission history

From: Yichong Xu [view email]
[v1] Mon, 6 Dec 2021 18:59:02 GMT (5124kb,D)
[v2] Tue, 14 Dec 2021 21:36:58 GMT (5125kb,D)
[v3] Wed, 4 May 2022 22:56:45 GMT (963kb,D)

Link back to: arXiv, form interface, contact.