Evidentiality-guided Generation for Knowledge-Intensive NLP Tasks

By Akari Asai, Matt Gardner, Hannaneh Hajishirzi. 2021

61 citations recorded by the source corpus.

Retrieval-augmented generation models have shown state-of-the-art performance across many knowledge-intensive NLP tasks such as open-domain question answering and fact verification. These models are trained to generate a final output given retrieved passages that can be irrelevant to an input query, leading to learning spurious cues or memorization. This work introduces a method to incorporate evidentiality of passages—whether a passage contains correct evidence to support the output—into training the generator. We introduce a multi-task learning framework to jointly generate the final output and predict the evidentiality of each passage. Furthermore, we introduce a new task-agnostic method for obtaining high-quality silver evidentiality labels, addressing the issues of gold evidentiality labels being unavailable in most domains. Our experiments on five datasets across three knowledge-intensive tasks show that our new evidentiality-guided generator significantly outperforms its direct counterpart on all of them, and advances the state of the art on three of them. Our analysis shows that multi-task learning and silver evidentiality mining play key roles. Our code is available at https://github.com/AkariAsai/evidentiality_qa

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Retrieval-Augmented Generation, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Retrieval-Augmented Generation

In Natural Language Processing

Citations in the corpus

Multi-Task Retrieval-Augmented Text Generation with Relevance Sampling (2022, background): “Our sampling is not limited to KILT and could be extended to other resources with a similar setup or by mining weakly-supervised relevance signals, as proposed by Asai et al. (2021), and filtering for example based on the confidence of the labelling model.”
When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories (2022, methodology): “Although increasing the context size often leads to performance gains (Izacard and Grave, 2021; Asai et al., 2022), we only use the top one retrieved paragraph for simplicity.”
When Not to Trust Language Models: Investigating Effectiveness and Limitations of Parametric and Non-Parametric Memories (2022, methodology): “Although increasing the context size often leads to performance gains (Izacard and Grave, 2021; Asai et al., 2022), we only use the top one retrieved paragraph for simplicity.”