Relational Memory-Augmented Language Models

Qi Liu, Dani Yogatama, Phil Blunsom


Abstract
We present a memory-augmented approach to condition an autoregressive language model on a knowledge graph. We represent the graph as a collection of relation triples and retrieve relevant relations for a given context to improve text generation. Experiments on WikiText-103, WMT19, and enwik8 English datasets demonstrate that our approach produces a better language model in terms of perplexity and bits per character. We also show that relational memory improves coherence, is complementary to token-based memory, and enables causal interventions. Our model provides a simple yet effective way to combine an autoregressive language model and a knowledge graph for more coherent and logical generation.
Anthology ID:
2022.tacl-1.32
Volume:
Transactions of the Association for Computational Linguistics, Volume 10
Month:
Year:
2022
Address:
Cambridge, MA
Editors:
Brian Roark, Ani Nenkova
Venue:
TACL
SIG:
Publisher:
MIT Press
Note:
Pages:
555–572
Language:
URL:
https://aclanthology.org/2022.tacl-1.32
DOI:
10.1162/tacl_a_00476
Bibkey:
Cite (ACL):
Qi Liu, Dani Yogatama, and Phil Blunsom. 2022. Relational Memory-Augmented Language Models. Transactions of the Association for Computational Linguistics, 10:555–572.
Cite (Informal):
Relational Memory-Augmented Language Models (Liu et al., TACL 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-1/2022.tacl-1.32.pdf