Thailand, Cambodia agree to build on ceasefire in talks in China's Yunnan Thailand and Cambodia plan to rebuild mutual trust and consolidate a ceasefire, Beijing says at the end of two days of talks in southwestern China, despite new accusations from the Thai military that its Cambodian counterparts are violating the truce with drone flights. The foreign ministers of Thailand and Cambodia met with the Chinese foreign minister in Yunnan province on Monday for the scheduled two days of talks aimed at ending weeks of fierce fighting along their border that has killed more than 100 people and displaced more than half a million civilians in both countries. As part of the deal, Thailand has agreed to return 18 captured Cambodian soldiers on Tuesday if the ceasefire, which took effect at noon (05:00 GMT) on Saturday, is fully observed. Speaking to reporters after the meeting, Thai Foreign Minister Sihasak Phuangketkeow said he believed the parties were "moving in a positive direction". "We haven't resolved everything, but I think we are making progress in the right direction, and we have to keep up the momentum," he said.

Cross-document Relation Extraction aims to predict the relation between target entities located in different documents. In this regard, the dominant models commonly retain useful information for relation prediction via bridge entities, which allows the model to elaborately capture the intrinsic interdependence between target entities. However, these studies ignore the non-bridge entities, each of which co-occurs with only one target entity and offers the semantic association between target entities for relation prediction. Besides, the commonly-used dataset--CodRED contains substantial NA instances, leading to the prediction bias during inference. To address these issues, in this paper, we propose a novel graph-based cross-document RE model with non-bridge entity enhancement and prediction debiasing. Specifically, we use a unified entity graph to integrate numerous non-bridge entities with target entities and bridge entities, modeling various associations between them, and then use a graph recurrent network to encode this graph. Finally, we introduce a novel debiasing strategy to calibrate the original prediction distribution. Experimental results on the closed and open settings show that our model significantly outperforms all baselines, including the GPT-3.5-turbo and InstructUIE, achieving state-of-the-art performance. Particularly, our model obtains 66.23% and 55.87% AUC points in the official leaderboard\footnote{\url{https://codalab.lisn.upsaclay.fr/competitions/3770#results}} under the two settings, respectively, ranking the first place in all submissions since December 2023. Our code is available at https://github.com/DeepLearnXMU/CoRE-NEPD.