This paper presents the system we developed for SemEval-2026 Task 1: Humor Generation. The task focuses on developing systems capable of generating genuinely humorous content under various constraints. In this work, we propose using a Retrieval-Augmented Generation approach to preprocess news headlines and obtain summaries of news content. Furthermore, we employ a unified humor generation mode to adapt to the two types of generation constraints. Finally, we conduct an internal-external voting process to produce the final optimal joke output. Our approach achieves competitive performance in this task: it ranks 1st (tied) among all participating teams in the Chinese track of Subtask A.

This paper describes our system to SemEval-2026 Task 3 Track A Subtask 1 on Dimensional Aspect Sentiment Regression (DimASR). We propose a lightweight and resource-efficient system built entirely on multilingual pre-trained encoders, without relying on LLMs or external corpora. We adopt joint multilingual and multi-domain training to facilitate cross-lingual transfer and alleviate data sparsity, introduce a bounded regression transformation that improves training stability while constraining predictions within the valid range, and employ an adaptive ensemble strategy via subset search to reduce prediction variance. Experimental results demonstrate that our system achieves strong and consistent performance, ranking 1st on zho-res, 2nd on zho-lap, and 3rd on jpn-hot, with all remaining datasets placed within the top half of participating teams.