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

Click-through rate (CTR) prediction plays as a core function module invarious personalized online services. The traditional ID-based models for CTRprediction take as inputs the one-hot encoded ID features of tabular modality,which capture the collaborative signals via feature interaction modeling. Butthe one-hot encoding discards the semantic information conceived in theoriginal feature texts. Recently, the emergence of Pretrained Language Models(PLMs) has given rise to another paradigm, which takes as inputs the sentencesof textual modality obtained by hard prompt templates and adopts PLMs toextract the semantic knowledge. However, PLMs generally tokenize the input textdata into subword tokens and ignore field-wise collaborative signals.Therefore, these two lines of research focus on different characteristics ofthe same input data (i.e., textual and tabular modalities), forming a distinctcomplementary relationship with each other. In this paper, we propose toconduct Fine-grained feature-level ALignment between ID-based Models andPretrained Language Models (FLIP) for CTR prediction. We design a novel jointreconstruction pretraining task for both masked language and tabular modeling.Specifically, the masked data of one modality (i.e., tokens or features) has tobe recovered with the help of the other modality, which establishes thefeature-level interaction and alignment via sufficient mutual informationextraction between dual modalities. Moreover, we propose to jointly finetunethe ID-based model and PLM for downstream CTR prediction tasks, thus achievingsuperior performance by combining the advantages of both models. Extensiveexperiments on three real-world datasets demonstrate that FLIP outperforms SOTAbaselines, and is highly compatible for various ID-based models and PLMs.

FLIP: Towards Fine-grained Alignment between ID-based Models and Pretrained Language Models for CTR Prediction