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

Vision-Language Pretraining (VLP) has shown impressive results on diversedownstream tasks by offline training on large-scale datasets. Regarding thegrowing nature of real-world data, such an offline training paradigm onever-expanding data is unsustainable, because models lack the continuallearning ability to accumulate knowledge constantly. However, most continuallearning studies are limited to uni-modal classification and existingmulti-modal datasets cannot simulate continual non-stationary data streamscenarios. To support the study of Vision-Language Continual Pretraining(VLCP), we first contribute a comprehensive and unified benchmark dataset P9Dwhich contains over one million product image-text pairs from 9 industries. Thedata from each industry as an independent task supports continual learning andconforms to the real-world long-tail nature to simulate pretraining on webdata. We comprehensively study the characteristics and challenges of VLCP, andpropose a new algorithm: Compatible momentum contrast with TopologyPreservation, dubbed CTP. The compatible momentum model absorbs the knowledgeof the current and previous-task models to flexibly update the modal feature.Moreover, Topology Preservation transfers the knowledge of embedding acrosstasks while preserving the flexibility of feature adjustment. The experimentalresults demonstrate our method not only achieves superior performance comparedwith other baselines but also does not bring an expensive training burden.Dataset and codes are available at https://github.com/KevinLight831/CTP.

CTP: Towards Vision-Language Continual Pretraining via Compatible Momentum Contrast and Topology Preservation