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

Mining the latent intentions from large volumes of natural language inputs isa key step to help data analysts design and refine Intelligent VirtualAssistants (IVAs) for customer service. To aid data analysts in this task wepresent Verint Intent Manager (VIM), an analysis platform that combinesunsupervised and semi-supervised approaches to help analysts quickly surfaceand organize relevant user intentions from conversational texts. For theinitial exploration of data we make use of a novel unsupervised andsemi-supervised pipeline that integrates the fine-tuning of high performinglanguage models, a distributed k-NN graph building method and communitydetection techniques for mining the intentions and topics from texts. Thefine-tuning step is necessary because pre-trained language models cannot encodetexts to efficiently surface particular clustering structures when the targettexts are from an unseen domain or the clustering task is not topic detection.For flexibility we deploy two clustering approaches: where the number ofclusters must be specified and where the number of clusters is detectedautomatically with comparable clustering quality but at the expense ofadditional computation time. We describe the application and deployment anddemonstrate its performance using BERT on three text mining tasks. Ourexperiments show that BERT begins to produce better task-aware representationsusing a labeled subset as small as 0.5% of the task data. The clusteringquality exceeds the state-of-the-art results when BERT is fine-tuned withlabeled subsets of only 2.5% of the task data. As deployed in the VIMapplication, this flexible clustering pipeline produces high quality results,improving the performance of data analysts and reducing the time it takes tosurface intentions from customer service data, thereby reducing the time ittakes to build and deploy IVAs in new domains.

A Semi-Supervised Deep Clustering Pipeline for Mining Intentions From Texts