Hao Fan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hao Fan has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Jia Wu
Macquarie University
6 shared publications - Jian Yang
School of Computing, Macquarie University, Sydney, NSW, Australia
5 shared publications - Xiaoxiao Ma
School of Computing, Macquarie University, Sydney, NSW, Australia
4 shared publications - Bingbing Xie
School of Information Management, Wuhan University, Wuhan, China
3 shared publications - Shan Xue
School of Computing, Macquarie University, Sydney, NSW, Australia
3 shared publications Ailimulati Yusupu
School Of Information Management, Wuhan University, Wuhan, 430072, China
1 shared publication
Latest Publications
- Open Access
- 1 February 2026
- Published by SAGE Publications in Digital Health
- Vol. 12
Understanding the influence of perceived intelligence and perceived anthropomorphism on users’ intention to adopt healthcare chatbots
Objectives: Healthcare chatbots have become essential tools for delivering health information and alleviating systemic pressures on medical resources. However, little research exists on how the AI attributes of healthcare chatbots affect users’ adoption intentions. Methods: Drawing on the stimulus-organism-response framework and tri-dimensional trusting beliefs theory, this study considers two AI attributes—perceived intelligence and perceived anthropomorphism—as stimuli and explores their effects on users’ functional perceptions (perceived usefulness, perceived ease of use, perceived fairness, perceived performance risk) and relational perceptions (competence trust, benevolence trust, integrity trust), and subsequently users’ intention to adopt healthcare chatbots. The research model was assessed using partial least squares structural equation modeling, based on survey data from 416 respondents in China. Results: The results demonstrate that both perceived intelligence and perceived anthropomorphism significantly influence users’ functional and relational perceptions. Perceived intelligence and perceived anthropomorphism enhance users’ intention to adopt healthcare chatbots through perceived usefulness, competence trust, and benevolence trust. Competence trust and benevolence trust mediate the effects of perceived intelligence and perceived anthropomorphism on users’ intention to adopt healthcare chatbots, while integrity trust did not significantly affect user adoption intention. Conclusions: The findings enhance the theoretical understanding of healthcare chatbot adoption and provide practical insights for their design and implementation.
- 1 February 2026
- Published by SAGE Publications in Digital Health
- Vol. 12
Open AccessUnderstanding the influence of perceived intelligence and perceived anthropomorphism on users’ intention to adopt healthcare chatbots
Objectives: Healthcare chatbots have become essential tools for delivering health information and alleviating systemic pressures on medical resources. However, little research exists on how the AI attributes of healthcare chatbots affect users’ adoption intentions. Methods: Drawing on the stimulus-organism-response framework and tri-dimensional trusting beliefs theory, this study considers two AI attributes—perceived intelligence and perceived anthropomorphism—as stimuli and explores their effects on users’ functional perceptions (perceived usefulness, perceived ease of use, perceived fairness, perceived performance risk) and relational perceptions (competence trust, benevolence trust, integrity trust), and subsequently users’ intention to adopt healthcare chatbots. The research model was assessed using partial least squares structural equation modeling, based on survey data from 416 respondents in China. Results: The results demonstrate that both perceived intelligence and perceived anthropomorphism significantly influence users’ functional and relational perceptions. Perceived intelligence and perceived anthropomorphism enhance users’ intention to adopt healthcare chatbots through perceived usefulness, competence trust, and benevolence trust. Competence trust and benevolence trust mediate the effects of perceived intelligence and perceived anthropomorphism on users’ intention to adopt healthcare chatbots, while integrity trust did not significantly affect user adoption intention. Conclusions: The findings enhance the theoretical understanding of healthcare chatbot adoption and provide practical insights for their design and implementation.
- Open Access
- 24 December 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15  (1) , 44467
Medical QA dialogue datasets in RAG systems performance evaluation and ChatGPT optimization
This study evaluates the effectiveness of Chinese doctor–patient dialogues as retrieval sources for Retrieval-Augmented Generation (RAG) in clinical question answering. Using ChatGPT-3.5 as a baseline and extending to GPT-4o and GPT-5, we compare multiple retrieval pipelines, including dense retrieval, Cross-Encoder reranking, Reciprocal Rank Fusion (RRF), and Cascade RRF→Rerank. Experimental results show that dialogue-based retrieval significantly improves generation quality relative to direct prompting (e.g., ROUGE-1-f: +12.6%, BERTScore_F1: +1.5%, p < 0.05). Among retrieval strategies, Rerank-only provides the best accuracy–latency balance, while the cascade pipeline introduces noise and yields no additional benefit. Under identical retrieval settings, GPT-4o achieves stronger automatic metrics and 4–5× lower latency, whereas GPT-5 receives slightly higher human preference scores (+ 0.08, p < 0.001), indicating a trade-off between efficiency and perceived coherence. Expert evaluation further confirms improvements in readability, accuracy, and authenticity (all p < 0.001). These findings highlight that data representation and metadata structure have a greater impact on RAG performance than retrieval algorithm complexity, offering practical guidance for reliable medical QA deployment.
- 24 December 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15  (1) , 44467
Open AccessMedical QA dialogue datasets in RAG systems performance evaluation and ChatGPT optimization
This study evaluates the effectiveness of Chinese doctor–patient dialogues as retrieval sources for Retrieval-Augmented Generation (RAG) in clinical question answering. Using ChatGPT-3.5 as a baseline and extending to GPT-4o and GPT-5, we compare multiple retrieval pipelines, including dense retrieval, Cross-Encoder reranking, Reciprocal Rank Fusion (RRF), and Cascade RRF→Rerank. Experimental results show that dialogue-based retrieval significantly improves generation quality relative to direct prompting (e.g., ROUGE-1-f: +12.6%, BERTScore_F1: +1.5%, p < 0.05). Among retrieval strategies, Rerank-only provides the best accuracy–latency balance, while the cascade pipeline introduces noise and yields no additional benefit. Under identical retrieval settings, GPT-4o achieves stronger automatic metrics and 4–5× lower latency, whereas GPT-5 receives slightly higher human preference scores (+ 0.08, p < 0.001), indicating a trade-off between efficiency and perceived coherence. Expert evaluation further confirms improvements in readability, accuracy, and authenticity (all p < 0.001). These findings highlight that data representation and metadata structure have a greater impact on RAG performance than retrieval algorithm complexity, offering practical guidance for reliable medical QA deployment.
- Open Access
- 14 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 39-47
A Study of Scientific Research User Switching Intention for AIGC application Platforms Under Human-AI Interaction: A Push-Pull-Mooring Perspective
- 14 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 39-47
Open AccessA Study of Scientific Research User Switching Intention for AIGC application Platforms Under Human-AI Interaction: A Push-Pull-Mooring Perspective
- 6 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 12  (5) , 3948-3961
Contrastive Multi-Knowledge Graph Learning for Fake News Detection
The widespread dissemination of fake news on social media platforms has significantly disrupted people's daily lives, causing substantial inconvenience and loss. Although fake news detection has captivated great research effort, most of the existing works either overlook the complementary ground knowledge, formulated in individual knowledge graphs (KGs) or ignore multiple aspects of knowledge in different knowledge graphs. In this paper, we propose a novel contrastive learning framework for investigating the complementary knowledge in different knowledge graphs, which achieves superior fake news detection performance. We first model three components within news articles (i.e., news content, entities, and topics) as a heterogeneous graph (HG). Subsequently, we leverage the valuable factual evidence among real-world entities, which is enclosed in two open-source knowledge graphs (Wikidata and DBpedia), and augment the heterogeneous graph into three augmented graphs. We then fuse knowledge-enhanced high-level news representations from each augmented graph and extract a unified news representation through our proposed contrastive multi-knowledge graph learning framework, which is ultimately utilized to identify fake news. We conduct extensive experiments on four public fake news datasets and demonstrate the superiority of our method over nine state-of-the-art baselines regarding five widely-used evaluation metrics (i.e., accuracy, precision, recall, F1-score, and ROC).
- 6 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 12  (5) , 3948-3961
Contrastive Multi-Knowledge Graph Learning for Fake News Detection
The widespread dissemination of fake news on social media platforms has significantly disrupted people's daily lives, causing substantial inconvenience and loss. Although fake news detection has captivated great research effort, most of the existing works either overlook the complementary ground knowledge, formulated in individual knowledge graphs (KGs) or ignore multiple aspects of knowledge in different knowledge graphs. In this paper, we propose a novel contrastive learning framework for investigating the complementary knowledge in different knowledge graphs, which achieves superior fake news detection performance. We first model three components within news articles (i.e., news content, entities, and topics) as a heterogeneous graph (HG). Subsequently, we leverage the valuable factual evidence among real-world entities, which is enclosed in two open-source knowledge graphs (Wikidata and DBpedia), and augment the heterogeneous graph into three augmented graphs. We then fuse knowledge-enhanced high-level news representations from each augmented graph and extract a unified news representation through our proposed contrastive multi-knowledge graph learning framework, which is ultimately utilized to identify fake news. We conduct extensive experiments on four public fake news datasets and demonstrate the superiority of our method over nine state-of-the-art baselines regarding five widely-used evaluation metrics (i.e., accuracy, precision, recall, F1-score, and ROC).
- 13 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 12  (2) , 682-694
Multiknowledge and LLM-Inspired Heterogeneous Graph Neural Network for Fake News Detection
The widespread diffusion of fake news has become a critical problem on dynamic social media worldwide, which requires effective strategies for fake news detection to alleviate its hazardous consequences for society. However, most recent efforts only focus on the features of news content and social context without realizing the benefits of large language models (LLMs) and multiple knowledge graphs (KGs), thus failing to improve detection capabilities further. To tackle this issue, we present a multiknowledge and LLM-inspired heterogeneous graph neural network for fake news detection (MiLk-FD), by combining KGs, LLMs, and graph neural networks (GNNs). Specifically, we first model news content as a heterogeneous graph (HG) containing news, entity, and topic nodes and then fuse the knowledge from three KGs to augment the factual basis of news articles. Meanwhile, we leverage TransE to initialize the knowledge features and employ LLaMa2-7B to obtain the initial feature vectors of news articles. After that, we utilize the devised HG transformer to learn news embeddings with specific feature distribution in high-dimensional spaces by aggregating neighborhood information according to metapaths. Finally, a classifier based on multilayer perceptron (MLP) is trained to predict each news article as fake or true. Through experiments, we demonstrate that our proposed framework surpasses ten baselines according to accuracy, precision, F1-score, recall, and ROC in four public real-world benchmarks (i.e., COVID-19, FakeNewsNet, PAN2020, Liar).
- 13 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 12  (2) , 682-694
Multiknowledge and LLM-Inspired Heterogeneous Graph Neural Network for Fake News Detection
The widespread diffusion of fake news has become a critical problem on dynamic social media worldwide, which requires effective strategies for fake news detection to alleviate its hazardous consequences for society. However, most recent efforts only focus on the features of news content and social context without realizing the benefits of large language models (LLMs) and multiple knowledge graphs (KGs), thus failing to improve detection capabilities further. To tackle this issue, we present a multiknowledge and LLM-inspired heterogeneous graph neural network for fake news detection (MiLk-FD), by combining KGs, LLMs, and graph neural networks (GNNs). Specifically, we first model news content as a heterogeneous graph (HG) containing news, entity, and topic nodes and then fuse the knowledge from three KGs to augment the factual basis of news articles. Meanwhile, we leverage TransE to initialize the knowledge features and employ LLaMa2-7B to obtain the initial feature vectors of news articles. After that, we utilize the devised HG transformer to learn news embeddings with specific feature distribution in high-dimensional spaces by aggregating neighborhood information according to metapaths. Finally, a classifier based on multilayer perceptron (MLP) is trained to predict each news article as fake or true. Through experiments, we demonstrate that our proposed framework surpasses ten baselines according to accuracy, precision, F1-score, recall, and ROC in four public real-world benchmarks (i.e., COVID-19, FakeNewsNet, PAN2020, Liar).
- Open Access
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 1272-1282
Heterogeneous Subgraph Transformer for Fake News Detection
Fake news is pervasive on social media, inflicting substantial harm on public discourse and societal well-being. We investigate the explicit structural information and textual features of news pieces by constructing a heterogeneous graph concerning the relations among news topics, entities, and content. Through our study, we reveal that fake news can be effectively detected in terms of the atypical heterogeneous subgraphs centered on them, which encapsulate the essential semantics and intricate relations between news elements. However, suffering from the heterogeneity, exploring such heterogeneous subgraphs remains an open problem. To bridge the gap, this work proposes a heterogeneous subgraph transformer HeteroSGT to exploit subgraphs in our constructed heterogeneous graph. In HeteroSGT, we first employ a pre-trained language model to derive both word-level and sentence-level semantics. Then the random walk with restart (RWR) is applied to extract subgraphs centered on each news, which are further fed to our proposed subgraph Transformer to quantify the authenticity. Extensive experiments on five real-world datasets demonstrate the superior performance of HeteroSGT over five baselines. Further case and ablation studies validate our motivation and demonstrate that performance improvement stems from our specially designed components.
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 1272-1282
Open AccessHeterogeneous Subgraph Transformer for Fake News Detection
Fake news is pervasive on social media, inflicting substantial harm on public discourse and societal well-being. We investigate the explicit structural information and textual features of news pieces by constructing a heterogeneous graph concerning the relations among news topics, entities, and content. Through our study, we reveal that fake news can be effectively detected in terms of the atypical heterogeneous subgraphs centered on them, which encapsulate the essential semantics and intricate relations between news elements. However, suffering from the heterogeneity, exploring such heterogeneous subgraphs remains an open problem. To bridge the gap, this work proposes a heterogeneous subgraph transformer HeteroSGT to exploit subgraphs in our constructed heterogeneous graph. In HeteroSGT, we first employ a pre-trained language model to derive both word-level and sentence-level semantics. Then the random walk with restart (RWR) is applied to extract subgraphs centered on each news, which are further fed to our proposed subgraph Transformer to quantify the authenticity. Extensive experiments on five real-world datasets demonstrate the superior performance of HeteroSGT over five baselines. Further case and ablation studies validate our motivation and demonstrate that performance improvement stems from our specially designed components.
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 508-521
On Fake News Detection with LLM Enhanced Semantics Mining
Xiaoxiao Ma, Yuchen Zhang, Kaize Ding, Jian Yang, Jia Wu, Hao Fan. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 508-521
Open AccessOn Fake News Detection with LLM Enhanced Semantics Mining
Xiaoxiao Ma, Yuchen Zhang, Kaize Ding, Jian Yang, Jia Wu, Hao Fan. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- 5 November 2023
- Published by Springer Nature
- p. 352-367
EmoKnow: Emotion- and Knowledge-Oriented Model for COVID-19 Fake News Detection
- 5 November 2023
- Published by Springer Nature
- p. 352-367
EmoKnow: Emotion- and Knowledge-Oriented Model for COVID-19 Fake News Detection
- Open Access
- 10 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Heterogeneous Graph Neural Network via Knowledge Relations for Fake News Detection
The proliferation of fake news in social media has been recognized as a severe problem for society, and substantial attempts have been devoted to fake news detection to alleviate the detrimental impacts. Knowledge graphs (KGs) comprise rich factual relations among real entities, which could be utilized as ground-truth databases and enhance fake news detection. However, most of the existing methods only leveraged natural language processing and graph mining techniques to extract features of fake news for detection and rarely explored the ground knowledge in knowledge graphs. In this work, we propose a novel Heterogeneous Graph Neural Network via Knowledge Relations for Fake News Detection (HGNNR4FD). The devised framework has four major components: 1) A heterogeneous graph (HG) built upon news content, including three types of nodes, i.e., news, entities, and topics, and their relations. 2) A KG that provides the factual basis for detecting fake news by generating embeddings via relations in the KG. 3) A novel attention-based heterogeneous graph neural network that can aggregate information from HG and KG, and 4) a fake news detector, which is capable of identifying fake news based on the news embeddings generated by HGNNR4FD. We further validate the performance of our method by comparison with seven state-of-art baselines and verify the effectiveness of the components through a thorough ablation analysis. From the results, we empirically demonstrate that our framework achieves superior results and yields improvement over the baselines regarding evaluation metrics of accuracy, precision, recall, and F1-score on four real-world datasets.
- 10 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Open AccessHeterogeneous Graph Neural Network via Knowledge Relations for Fake News Detection
The proliferation of fake news in social media has been recognized as a severe problem for society, and substantial attempts have been devoted to fake news detection to alleviate the detrimental impacts. Knowledge graphs (KGs) comprise rich factual relations among real entities, which could be utilized as ground-truth databases and enhance fake news detection. However, most of the existing methods only leveraged natural language processing and graph mining techniques to extract features of fake news for detection and rarely explored the ground knowledge in knowledge graphs. In this work, we propose a novel Heterogeneous Graph Neural Network via Knowledge Relations for Fake News Detection (HGNNR4FD). The devised framework has four major components: 1) A heterogeneous graph (HG) built upon news content, including three types of nodes, i.e., news, entities, and topics, and their relations. 2) A KG that provides the factual basis for detecting fake news by generating embeddings via relations in the KG. 3) A novel attention-based heterogeneous graph neural network that can aggregate information from HG and KG, and 4) a fake news detector, which is capable of identifying fake news based on the news embeddings generated by HGNNR4FD. We further validate the performance of our method by comparison with seven state-of-art baselines and verify the effectiveness of the components through a thorough ablation analysis. From the results, we empirically demonstrate that our framework achieves superior results and yields improvement over the baselines regarding evaluation metrics of accuracy, precision, recall, and F1-score on four real-world datasets.
- Open Access
- 11 April 2020
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 17  (8) , 2625
Mapping the Scientific Research on Healthcare Workers’ Occupational Health: A Bibliometric and Social Network Analysis
In the last few years, the occupational health (OH) of healthcare workers (HCWs) has been shown increasing concern by both health departments and researchers. This study aims to provide academics with quantitative and qualitative analysis of healthcare workers’ occupational health (HCWs+OH) field in a joint way. Based on 402 papers published from 1992 to 2019, we adopted the approaches of bibliometric and social network analysis (SNA) to map and quantify publication years, research area distribution, international collaboration, keyword co-occurrence frequency, hierarchical clustering, highly cited articles and cluster timeline visualization. In view of the results, several hotspot clusters were identified, namely: physical injuries, workplace, mental health; occupational hazards and diseases, infectious factors; community health workers and occupational exposure. As for citations, we employed document co-citation analysis to detect trends and identify seven clusters, namely tuberculosis (TB), strength training, influenza, healthcare worker (HCW), occupational exposure, epidemiology and psychological. With the visualization of cluster timeline, we detected that the earliest research cluster was occupational exposure, then followed by epidemiology and psychological; however, TB, strength training and influenza appeared to gain more attention in recent years. These findings are presumed to offer researchers, public health practitioners a comprehensive understanding of HCWs+OH research.
- 11 April 2020
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 17  (8) , 2625
Open AccessMapping the Scientific Research on Healthcare Workers’ Occupational Health: A Bibliometric and Social Network Analysis
In the last few years, the occupational health (OH) of healthcare workers (HCWs) has been shown increasing concern by both health departments and researchers. This study aims to provide academics with quantitative and qualitative analysis of healthcare workers’ occupational health (HCWs+OH) field in a joint way. Based on 402 papers published from 1992 to 2019, we adopted the approaches of bibliometric and social network analysis (SNA) to map and quantify publication years, research area distribution, international collaboration, keyword co-occurrence frequency, hierarchical clustering, highly cited articles and cluster timeline visualization. In view of the results, several hotspot clusters were identified, namely: physical injuries, workplace, mental health; occupational hazards and diseases, infectious factors; community health workers and occupational exposure. As for citations, we employed document co-citation analysis to detect trends and identify seven clusters, namely tuberculosis (TB), strength training, influenza, healthcare worker (HCW), occupational exposure, epidemiology and psychological. With the visualization of cluster timeline, we detected that the earliest research cluster was occupational exposure, then followed by epidemiology and psychological; however, TB, strength training and influenza appeared to gain more attention in recent years. These findings are presumed to offer researchers, public health practitioners a comprehensive understanding of HCWs+OH research.