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

Multimedia recommendation aims to predict users' future behaviors based onhistorical behavioral data and item's multimodal information. However, noiseinherent in behavioral data, arising from unintended user interactions withuninteresting items, detrimentally impacts recommendation performance.Recently, diffusion models have achieved high-quality information generation,in which the reverse process iteratively infers future information based on thecorrupted state. It meets the need of predictive tasks under noisy conditions,and inspires exploring their application to predicting user behaviors.Nonetheless, several challenges must be addressed: 1) Classical diffusionmodels require excessive computation, which does not meet the efficiencyrequirements of recommendation systems. 2) Existing reverse processes aremainly designed for continuous data, whereas behavioral information is discretein nature. Therefore, an effective method is needed for the generation ofdiscrete behavioral information.  To tackle the aforementioned issues, we propose a Light Diffusion model forMultimedia Recommendation. First, to reduce computational complexity, wesimplify the formula of the reverse process, enabling one-step inferenceinstead of multi-step inference. Second, to achieve effective behavioralinformation generation, we propose a novel Conditional neural Network. It mapsthe discrete behavior data into a continuous latent space, and generatesbehaviors with the guidance of collaborative signals and user multimodalpreference. Additionally, considering that completely clean behavior data isinaccessible, we introduce a soft behavioral reconstruction constraint duringmodel training, facilitating behavior prediction with noisy data. Empiricalstudies conducted on three public datasets demonstrate the effectiveness ofLD4MRec.

LD4MRec: Simplifying and Powering Diffusion Model for Multimedia Recommendation