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

Retrieval-based conversational systems learn to rank response candidates fora given dialogue context by computing the similarity between their vectorrepresentations. However, training on a single textual form of the multi-turncontext limits the ability of a model to learn representations that generalizeto natural perturbations seen during inference. In this paper we propose aframework that incorporates augmented versions of a dialogue context into thelearning objective. We utilize contrastive learning as an auxiliary objectiveto learn robust dialogue context representations that are invariant toperturbations injected through the augmentation method. We experiment with fourbenchmark dialogue datasets and demonstrate that our framework combines wellwith existing augmentation methods and can significantly improve over baselineBERT-based ranking architectures. Furthermore, we propose a novel dataaugmentation method, ConMix, that adds token level perturbations throughstochastic mixing of tokens from other contexts in the batch. We show that ourproposed augmentation method outperforms previous data augmentation approaches,and provides dialogue representations that are more robust to commonperturbations seen during inference.

DialAug: Mixing up Dialogue Contexts in Contrastive Learning for Robust Conversational Modeling