FactKB: Generalizable Factuality Evaluation using Language Models Enhanced with Factual Knowledge
This addresses the need for robust and generalizable factuality evaluation in summarization systems, particularly for handling entity and relation errors across domains, though it is incremental in its approach.
The authors tackled the problem of evaluating factual consistency in automatically generated summaries, which is prone to errors in new domains, by proposing FactKB, a language model enhanced with factual knowledge from external knowledge bases. The result is state-of-the-art performance on in-domain news and out-of-domain scientific benchmarks, with improved detection of entity and relation errors.
Evaluating the factual consistency of automatically generated summaries is essential for the progress and adoption of reliable summarization systems. Despite recent advances, existing factuality evaluation models are not robust, being especially prone to entity and relation errors in new domains. We propose FactKB, a simple new approach to factuality evaluation that is generalizable across domains, in particular with respect to entities and relations. FactKB is based on language models pretrained using facts extracted from external knowledge bases. We introduce three types of complementary factuality pretraining objectives based on direct entity facts, facts grounded in auxiliary knowledge about entities, and facts constructed compositionally through knowledge base walks. The resulting factuality evaluation model achieves state-of-the-art performance on two in-domain news summarization benchmarks as well as on three out-of-domain scientific literature datasets. Further analysis of FactKB shows improved ability to detect erroneous entities and relations in summaries and is robust and generalizable across domains.