English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

While large pre-trained language models are powerful, their predictions oftenlack logical consistency across test inputs. For example, a state-of-the-artMacaw question-answering (QA) model answers 'Yes' to 'Is a sparrow a bird?' and'Does a bird have feet?' but answers 'No' to 'Does a sparrow have feet?'. Toaddress this failure mode, we propose a framework, Consistency Correctionthrough Relation Detection, or ConCoRD, for boosting the consistency andaccuracy of pre-trained NLP models using pre-trained natural language inference(NLI) models without fine-tuning or re-training. Given a batch of test inputs,ConCoRD samples several candidate outputs for each input and instantiates afactor graph that accounts for both the model's belief about the likelihood ofeach answer choice in isolation and the NLI model's beliefs about pair-wiseanswer choice compatibility. We show that a weighted MaxSAT solver canefficiently compute high-quality answer choices under this factor graph,improving over the raw model's predictions. Our experiments demonstrate thatConCoRD consistently boosts accuracy and consistency of off-the-shelfclosed-book QA and VQA models using off-the-shelf NLI models, notablyincreasing accuracy of LXMERT on ConVQA by 5% absolute. Seehttps://ericmitchell.ai/emnlp-2022-concord/ for code and data.

Enhancing Self-Consistency and Performance of Pre-Trained Language Models through Natural Language Inference