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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Machine-learning based recommender systems(RSs) has become an effective meansto help people automatically discover their interests. Existing models oftenrepresent the rich information for recommendation, such as items, users, andcontexts, as embedding vectors and leverage them to predict users' feedback. Inthe view of causal analysis, the associations between these embedding vectorsand users' feedback are a mixture of the causal part that describes why an itemis preferred by a user, and the non-causal part that merely reflects thestatistical dependencies between users and items, for example, the exposuremechanism, public opinions, display position, etc. However, existing RSs mostlyignored the striking differences between the causal parts and non-causal partswhen using these embedding vectors. In this paper, we propose a model-agnosticframework named IV4Rec that can effectively decompose the embedding vectorsinto these two parts, hence enhancing recommendation results. Specifically, wejointly consider users' behaviors in search scenarios and recommendationscenarios. Adopting the concepts in causal analysis, we embed users' searchbehaviors as instrumental variables (IVs), to help decompose original embeddingvectors in recommendation, i.e., treatments. IV4Rec then combines the two partsthrough deep neural networks and uses the combined results for recommendation.IV4Rec is model-agnostic and can be applied to a number of existing RSs such asDIN and NRHUB. Experimental results on both public and proprietary industrialdatasets demonstrate that IV4Rec consistently enhances RSs and outperforms aframework that jointly considers search and recommendation.

A Model-Agnostic Causal Learning Framework for Recommendation using Search Data