Abstract | ||
---|---|---|
Past research has demonstrated that large neural language models (LMs) encode surprising amounts of factual information: however, augmenting or modifying this information requires modifying a corpus and retraining, which is computationally expensive. To address this problem, we develop a neural LM that includes an interpretable neuro-symbolic KB in the form of a “fact memory”. Each element of the fact memory is formed from a triple of vectors, where each vector corresponds to a KB entity or relation. Our LM improves performance on knowledge-intensive question-answering tasks, sometimes dramatically, including a 27 point increase in one setting of WebQuestionsSP over a state-of-the-art open-book model, despite using 5% of the parameters. Most interestingly, we demonstrate that the model can be modified, without any re-training, by updating the fact memory. |
Year | Venue | DocType |
---|---|---|
2021 | NAACL-HLT | Conference |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Patrick Verga | 1 | 97 | 9.11 |
Sun, Haitian | 2 | 13 | 4.28 |
Livio B. Soares | 3 | 0 | 0.68 |
William W. Cohen | 4 | 10178 | 1243.74 |