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

Perceiving multi-modal information and fulfilling dialogues with humans is along-term goal of artificial intelligence. Pre-training is commonly regarded asan effective approach for multi-modal dialogue. However, due to the limitedavailability of multi-modal dialogue data, there is still scarce research onmulti-modal dialogue pre-training. Yet another intriguing challenge emergesfrom the encompassing nature of multi-modal dialogue, which involves variousmodalities and tasks. Moreover, new forms of tasks may arise at unpredictablepoints in the future. Hence, it is essential for designed multi-modal dialoguemodels to possess sufficient flexibility to adapt to such scenarios. This paperproposes \textbf{PaCE}, a unified, structured, compositional multi-modaldialogue pre-training framework. It utilizes a combination of severalfundamental experts to accommodate multiple dialogue-related tasks and can bepre-trained using limited dialogue and extensive non-dialogue multi-modal data.Furthermore, we propose a progressive training method where old experts fromthe past can assist new experts, facilitating the expansion of theircapabilities. Experimental results demonstrate that PaCE achievesstate-of-the-art results on eight multi-modal dialog benchmarks.

PaCE: Unified Multi-modal Dialogue Pre-training with Progressive and Compositional Experts