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

Generative commonsense question answering (GenCQA) is a task of automaticallygenerating a list of answers given a question. The answer list is required tocover all reasonable answers. This presents the considerable challenges ofproducing diverse answers and ranking them properly. Incorporating a variety ofclosely-related background knowledge into the encoding of questions enables thegeneration of different answers. Meanwhile, learning to distinguish positiveanswers from negative ones potentially enhances the probabilistic estimation ofplausibility, and accordingly, the plausibility-based ranking. Therefore, wepropose a Knowledge Enhancement and Plausibility Ranking (KEPR) approachgrounded on the Generate-Then-Rank pipeline architecture. Specifically, weexpand questions in terms of Wiktionary commonsense knowledge of keywords, andreformulate them with normalized patterns. Dense passage retrieval is utilizedfor capturing relevant knowledge, and different PLM-based (BART, GPT2 and T5)networks are used for generating answers. On the other hand, we develop anELECTRA-based answer ranking model, where logistic regression is conductedduring training, with the aim of approximating different levels of plausibilityin a polar classification scenario. Extensive experiments on the benchmarkProtoQA show that KEPR obtains substantial improvements, compared to the strongbaselines. Within the experimental models, the T5-based GenCQA with KEPRobtains the best performance, which is up to 60.91% at the primary canonicalmetric Inc@3. It outperforms the existing GenCQA models on the currentleaderboard of ProtoQA.

KEPR: Knowledge Enhancement and Plausibility Ranking for Generative Commonsense Question Answering