Ma Jingkun, a master’s student in the Faculty of Science and Technology (FST) at the University of Macau (UM), won a first prize in the final round of the 17th Pan-Pearl River Delta+ University IT Project Competition for his project ‘Yu Sheng: Human-in-Loop Classical Chinese Poetry Generation System’. He won a third prize in the 10th ‘Sharing Cup’ Innovation Competition of Science and Technology Resources Sharing Service. The system uses artificial intelligence technology to simulate and assist in the creation of traditional Chinese poetry and provide a poetry creation platform that integrates automatic poetry generation and refinement for poetry lovers.
Ma developed the system using cutting-edge natural language processing and deep learning technologies under the guidance of his supervisor, Wong Fai, professor in the Natural Language Processing & Portuguese-Chinese Machine Translation Laboratory (NLP2CT) at UM. The system provides various customisation options to meet users’ diverse needs. It has a wide range of custom configuration functions for poetry, covering poetic features such as words, sentences, tones, and rhyme. In addition, it can provide dynamic and static two-dimensional refinement functions according to different scene settings. These include automatic refinement, inspiration expansion, poetry import, and analysis, enabling the system to cover the entire automated process of ‘generation + refinement’ of poetry. Based on large-scale pre-trained language models, the system is designed and optimised for relevant deep learning algorithms according to task scenarios. It also applies model acceleration technology to provide efficient and high-quality poetry creation and optimisation services.
In addition to the aforementioned awards, Ma also received a first prize in the IEEE Industrial Electronics Society (IES) Undergraduate Contest and a third prize in the IEEE Macau Student Branch Project Competition. He also entered the FST Dean’s Final Year Project List for the system. His paper on the system was also accepted by the European Chapter of the Association for Computational Linguistics (EACL 2023). For those who are interested in using the system, please visit https://yusheng.cis.um.edu.mo/.
UM is committed to cultivating high-calibre talent with global perspectives and innovation. The awards and academic achievements attained by Ma’s ‘Yu Sheng: Human-in-Loop Classical Chinese Poetry Generation System’ reflects the university’s efforts and achievements in the field of natural language processing and deep learning. UM hopes that in the future, more students will play a role in technological innovation at home and abroad and contribute to the development of science and technology in Macao and the country.
澳門大學科技學院碩士生馬景坤開發的“語笙•智慧詩歌創作系統”在第17屆“泛珠三角+大學生計算機作品賽總決賽”中獲得一等獎,並在10屆“共享盃” 科技資源共享服務創新大賽獲得三等獎。該系統利用人工智能技術模擬和輔助中華傳統詩歌的創作,為詩歌愛好者提供一個集自動生成和潤色功能於一身的創作平台。
在澳大自然語言處理與中葡機器翻譯實驗室(NLP2CT)的黃輝教授指導下,馬景坤運用前沿的自然語言處理與深度學習技術開發系統。該系統能根據使用者的多元化需求提供豐富選項,例如設定言、句、平仄、韻部等詩歌特徵,亦可按不同的場景設定提供動態和靜態的二維潤色功能,包括自動潤色、靈感擴充、詩歌導入、分析等,實現詩歌“生成+潤色”的自動化流程。該系統以大型預訓練語言模型為基礎,結合了按特定任務場景設計和優化的深度學習演算法,加上模型加速技術,提供高效而優質的詩歌創作和優化服務。
除上述獎項,該系統亦在“電氣與電子工程師協會澳門工業電子學會本科生項目比賽”中獲得一等獎、在“電氣與電子工程師協會澳門學生分會項目比賽”中獲得三等獎,並獲澳大科技學院列入院長榮譽榜。相關學術論文亦獲計算語言學協會歐洲分會會議(EACL 2023)錄取。歡迎有興趣人士使用系統:https://yusheng.cis.um.edu.mo/。
澳大一直致力培養具備國際視野和創新能力的高素質人才。“語笙•智慧詩歌創作系統”取得的獎項和學術成就,體現了澳大在自然語言處理和深度學習方面的努力和成果。澳大期望未來有更多學生能夠走上海內外的科技創新舞台,為澳門和國家的科技發展貢獻力量。