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

Abductive and counterfactual reasoning, core abilities of everyday humancognition, require reasoning about what might have happened at time t, whileconditioning on multiple contexts from the relative past and future. However,simultaneous incorporation of past and future contexts using generativelanguage models (LMs) can be challenging, as they are trained either tocondition only on the past context or to perform narrowly scopedtext-infilling. In this paper, we propose DeLorean, a new unsupervised decodingalgorithm that can flexibly incorporate both the past and future contexts usingonly off-the-shelf, left-to-right language models and no supervision. The keyintuition of our algorithm is incorporating the future throughback-propagation, during which, we only update the internal representation ofthe output while fixing the model parameters. By alternating between forwardand backward propagation, DeLorean can decode the output representation thatreflects both the left and right contexts. We demonstrate that our approach isgeneral and applicable to two nonmonotonic reasoning tasks: abductive textgeneration and counterfactual story revision, where DeLorean outperforms arange of unsupervised and some supervised methods, based on automatic and humanevaluation.

Back to the Future: Unsupervised Backprop-based Decoding for Counterfactual and Abductive Commonsense Reasoning