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

Although pre-trained language models~(PLMs) have recently advanced theresearch progress in mathematical reasoning, they are not specially designed asa capable multi-task solver, suffering from high cost for multi-task deployment(\eg a model copy for a task) and inferior performance on complex mathematicalproblems in practical applications. To address these issues, in this paper, wepropose \textbf{JiuZhang~2.0}, a unified Chinese PLM specially for multi-taskmathematical problem solving. Our idea is to maintain a moderate-sized modeland employ the \emph{cross-task knowledge sharing} to improve the modelcapacity in a multi-task setting. Specially, we construct aMixture-of-Experts~(MoE) architecture for modeling mathematical text, so as tocapture the common mathematical knowledge across tasks. For optimizing the MoEarchitecture, we design \emph{multi-task continual pre-training} and\emph{multi-task fine-tuning} strategies for multi-task adaptation. Thesetraining strategies can effectively decompose the knowledge from the task dataand establish the cross-task sharing via expert networks. In order to furtherimprove the general capacity of solving different complex tasks, we leveragelarge language models~(LLMs) as complementary models to iteratively refine thegenerated solution by our PLM, via in-context learning. Extensive experimentshave demonstrated the effectiveness of our model.

JiuZhang 2.0: A Unified Chinese Pre-trained Language Model for Multi-task Mathematical Problem Solving