太阳成集团tyc9728·(中国)唯一官方网站

Contact Information

School of Information Management,
Wuhan University,
Wuhan, Hubei Province,
P.R.China. 430072

fuling@whu.edu.cn

SIM Teachers and Students Participated in iConference2024, and Dean Wei Lu Gave a Keynote Speech

2024-05-14 11:19:15

Recently, iConference2024, the 19th annual meeting of iSchools, co-sponsored by Jilin University and University of Tsukuba, was held in Changchun. Feicheng Ma, Senior Professor of Humanities and Social Sciences at Wuhan University, and Chuanfu Chen, Senior Professor of Humanities and Social Sciences at Wuhan University, attended the opening ceremony, delivered opening speeches and presented the award of best short research paper respectively. Wei Lu, Dean of our School, was invited to give a keynote speech.

Feicheng Ma expressed in his speech that the wave of digitalization and intelligence is developing rapidly in the new era. A new round of scientific and technological revolution represented by emerging technologies such as big data, artificial intelligence and Internet of Things is reshaping the production and lifestyle of human society and promoting the transformation and upgrading of intelligence in all fields. It is believed that the status quo, challenges and future development of information management and information science in the digital environment, the development of iSchools and many other frontier propositions will be fully discussed via iConference, a platform that gathers global wisdom.

In the report entitled Originality Evaluation based on the Processing and Understanding of Scientific Papers with AI, Wei Lu introduced the group's previous research work on using AI technology to process and understand scientific texts, and reported two latest studies.

In the first study, the group analyzed the feasibility of using large language model to evaluate originality in zero sample learning. The performance of originality evaluation was tested on different large language models by designing a novel instruction and constructing two new evaluation datasets. It is found that the large language model can be used as a qualified reviewer to some extent, but its evaluation performance is not perfect.

In the second study, the group constructed a new scoring dataset, and fine-tuned the large language model by QLoRA algorithm, which significantly improved the evaluation performance of the large language model. Specifically, the model built by his group can effectively distinguish new articles published in journals with different impact factors according to the research content merely.

Wei Lu also introduced and demonstrated the newly developed originality evaluation system, which can generate originality score, originality category and originality description to evaluate scientific articles.

The theme of iConference2024 is Wisdom, Well-being, Win-win, aiming at exploring the formation and development of new topics in the field of data, information and knowledge in the digital age. The conference was held in a combination of virtual and onsite meeting, and about 300 experts and scholars from global information science and related fields attended the onsite meeting.

8 full research papers, 5 English short research papers, 4 English posters, 6 Chinese papers and 3 Chinese posters were accepted in iConference2024. Those papers and posters’ first authors are our School’s teachers or students.

Professor Xiaoguang Wang presided over the Chinese keynote speech, Professor Lihong Zhou, Treasurer of iSchools, presided over the keynote speech at the opening ceremony, Professor Lu An participated in the Chinese forum as commentator, and Associate Professor Jin Mao’s Chinese paper Research on Early Identification Methods of Emerging Topics Based on Weak Signal 3D Framework was awarded the Best Chinese Paper.

Baidu
sogou