A5101541888- Multimodal Machine Learning Applications
- Coastal and Marine Management
- Social Media and Politics
- Digital Storytelling and Education
- Video Analysis and Summarization
- Subtitles and Audiovisual Media
- Energy Load and Power Forecasting
- Fault Detection and Control Systems
- Misinformation and Its Impacts
- Handwritten Text Recognition Techniques
- Neural Networks and Applications
- Media Studies and Communication
- Digital Humanities and Scholarship
- Hong Kong and Taiwan Politics
- Air Quality Monitoring and Forecasting
- Forecasting Techniques and Applications
- Hydrological Forecasting Using AI
- Data Visualization and Analytics
- Internet Traffic Analysis and Secure E-voting
Hong Kong Baptist University
2021-2022
Guangdong University of Foreign Studies
2018
Zhejiang Ocean University
2010-2011
Although studies have investigated cyber-rumoring previous to the pandemic, little research has been undertaken study rumors and rumor-corrections during COVID-19 (coronavirus disease 2019) pandemic. Drawing on prior about how online stories become viral, this will fill that gap by investigating retransmission of corrective messages Sina Weibo, largest most popular microblogging site in China. This examines impact rumor types, content attributes (including frames, emotion, rationality),...
Despite significant progress in generative AI, comprehensive evaluation remains challenging because of the lack effective metrics and standardized benchmarks. For instance, widely-used CLIPScore measures alignment between a (generated) image text prompt, but it fails to produce reliable scores for complex prompts involving compositions objects, attributes, relations. One reason is that encoders CLIP can notoriously act as "bag words", conflating such "the horse eating grass" with grass...
While text-to-visual models now produce photo-realistic images and videos, they struggle with compositional text prompts involving attributes, relationships, higher-order reasoning such as logic comparison. In this work, we conduct an extensive human study on GenAI-Bench to evaluate the performance of leading image video generation in various aspects generation. We also compare automated evaluation metrics against our collected ratings find that VQAScore -- a metric measuring likelihood VQA...
Purpose This study seeks to establish a new framework for categorizing incivility, differentiating between explicit and implicit forms, investigate their respective abilities proliferate mobilize conversations, along with behavioral outcomes in various social contexts. Design/methodology/approach Employing computational techniques, this research analyzed 10,145 protest-related threads from the HK Golden Forum, prominent online discussion board Hong Kong. Findings Our analysis revealed...
Vision-language models (VLMs) have made significant progress in recent visual-question-answering (VQA) benchmarks that evaluate complex visio-linguistic reasoning. However, are these truly effective? In this work, we show VLMs still struggle with natural images and questions humans can easily answer, which term adversarial samples. We also find it surprisingly easy to generate VQA samples from image-text corpora using off-the-shelf like CLIP ChatGPT. propose a semi-automated approach collect...