Zhejun Zhang

Zhejun Zhang

Auto-generated profile, based on publicly available information

Zhejun Zhang

School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
2 February 2024 in AI
Verified
Journal Article
Open access
PDF file
ZZZhejun Zhang
Zhejun Zhang
HCHuiying Chen
Huiying Chen
RHRuonan Huang
Ruonan Huang
LZLihong Zhu
Lihong Zhu
SMShengling Ma
Shengling Ma
Wei LiuWei Liu
Wei Liu
This study introduces a novel tool for classifying user needs in user experience (UX) design, specifically tailored for beginners, with potential applications in education. The tool employs the Kano model, text analysis, and deep learning to classify user needs efficiently into four categories. The data for the study were collected through interviews and web crawling, yielding 19 user needs from Generation Z users (born between 1995 and 2009) of LEGO toys (Billund, Denmark). These needs were then categorized into must-be, one-dimensional, attractive, and indifferent needs through a Kano-based questionnaire survey. A dataset of over 3000 online comments was created through preprocessing and annotating, which was used to train and evaluate seven deep learning models. The most effective model, the Recurrent Convolutional Neural Network (RCNN), was employed to develop a graphical text classification tool that accurately outputs the corresponding category and probability of user input text according to the Kano model. A usability test compared the tool’s performance to the traditional affinity diagram method. The tool outperformed the affinity diagram method in six dimensions and outperformed three qualities of the User Experience Questionnaire (UEQ), indicating a superior UX. The tool also demonstrated a lower perceived workload, as measured using the NASA Task Load Index (NASA-TLX), and received a positive Net Promoter Score (NPS) of 23 from the participants. These findings underscore the potential of this tool as a valuable educational resource in UX design courses. It offers students a more efficient and engaging and less burdensome learning experience while seamlessly integrating artificial intelligence into UX design education. This study provides UX design beginners with a practical and intuitive tool, facilitating a deeper understanding of user needs and innovative design strategies.

2 January 2025 in Scientific Reports
Verified
Journal Article
Open access
PDF file
ZZZhejun Zhang
Zhejun Zhang
SGShaoting Guo
Shaoting Guo
WZWenqing Zhou
Wenqing Zhou
YLYingying Luo
Yingying Luo
YZYingqi Zhu
Yingqi Zhu
LLLei Li
Lei Li
Deep language models (DLMs) have exhibited remarkable language understanding and generation capabilities, prompting researchers to explore the similarities between their internal mechanisms and human language cognitive processing. This study investigated the representational similarity (RS) between the abstractive summarization (ABS) models and the human brain and its correlation to the performance of ABS tasks. Specifically, representational similarity analysis (RSA) was used to measure the similarity between the representational patterns (RPs) of the BART, PEGASUS, and T5 models’ hidden layers and the human brain’s language RPs under different spatiotemporal conditions. Layer-wise ablation manipulation, including attention ablation and noise addition was employed to examine the hidden layers’ effect on model performance. The results demonstrate that as the depth of hidden layers increases, the models’ text encoding becomes increasingly similar to the human brain’s language RPs. Manipulating deeper layers leads to more substantial decline in summarization performance compared to shallower layers, highlighting the crucial role of deeper layers in integrating essential information. Notably, the study confirms the hypothesis that the hidden layers exhibiting higher similarity to human brain activity play a more critical role in model performance, with their correlations reaching statistical significance even after controlling for perplexity. These findings deepen our understanding of the cognitive mechanisms underlying language representations in DLMs and their neural correlates, potentially providing insights for optimizing and improving language models by aligning them with the human brain’s language-processing mechanisms.

29 May 2025 in IEEE Transactions on Audio, Speech and Language Processing
Verified
Journal Article
Closed access
YZYingqi Zhu
Yingqi Zhu
YLYiyang Li
Yiyang Li
LLLei Li
Lei Li
DHDingxin Hu
Dingxin Hu
XHXueyi Hao
Xueyi Hao
Natalia VanetikNatalia Vanetik
Natalia Vanetik
Improving factual consistency in abstractive summarization has been a focus of recent research. One promising approach is the post-editing method. However, previous works have yet to make sufficient use of factual factors in summaries and suffer from the negative effect of the training datasets. In this paper, we first propose a novel factual error correction model FactCloze based on a conditional-generation cloze task. FactCloze can construct the causality among factual factors while being able to determine whether the blank can be answered. Then, we propose a data distillation method to generate a more faithful summarization dataset SummDSC via multiple-dimensional evaluation. We validate our method on both non-LLM and LLM-generated datasets. Besides BART and T5, we implement FactCloze using DeepSeek prompt. Finally, we examine the differences between LLM-based and traditional evaluation metrics for factual error correction.

19 January 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ZZZhejun Zhang
Zhejun Zhang
YZYingqi Zhu
Yingqi Zhu
YZYubo Zheng
Yubo Zheng
YLYingying Luo
Yingying Luo
HSHengyi Shao
Hengyi Shao
LLLei Li
Lei Li
(1) Background: Information overload challenges decision-making in the Industry 4.0 era. While Natural Language Processing (NLP), especially Automatic Text Summarization (ATS), offers solutions, issues with factual accuracy persist. This research bridges cognitive neuroscience and NLP, aiming to improve model interpretability. (2) Methods: This research examined four fact extraction techniques: dependency relation, named entity recognition, part-of-speech tagging, and TF-IDF, in order to explore their correlation with human EEG signals. Representational Similarity Analysis (RSA) was applied to gauge the relationship between language models and brain activity. (3) Results: Named entity recognition showed the highest sensitivity to EEG signals, marking the most significant differentiation between factual and non-factual words with a score of −0.99. The dependency relation followed with −0.90, while part-of-speech tagging and TF-IDF resulted in 0.07 and −0.52, respectively. Deep language models such as GloVe, BERT, and GPT-2 exhibited noticeable influences on RSA scores, highlighting the nuanced interplay between brain activity and these models. (4) Conclusions: Our findings emphasize the crucial role of named entity recognition and dependency relations in fact extraction and demonstrate the independent effects of different models and TOIs on RSA scores. These insights aim to refine algorithms to reflect human text processing better, thereby enhancing ATS models’ factual integrity.

11 February 2026 in Biomedical Signal Processing and Control
Verified
Journal Article
Closed access
LDLiang Dong
Liang Dong
HSHengyi Shao
Hengyi Shao
ZZZhejun Zhang
Zhejun Zhang
YZYingqi Zhu
Yingqi Zhu
SGShaoting Guo
Shaoting Guo
LLLei Li
Lei Li

30 March 2026 in Neural Networks
Verified
Journal Article
Closed access
ZZZhejun Zhang
Zhejun Zhang
WZWenqing Zhou
Wenqing Zhou
SZShuo Zhang
Shuo Zhang
XLXinhang Li
Xinhang Li
LZLin Zhang
Lin Zhang
LLLei Li
Lei Li