Members of Associate Professor Dong Wang's research group, the Social Sensing Lab, will present papers at the 2021 IEEE International Conference on Big Data (IEEE BigData 2021), which will be held virtually from December 15-18.
PhD student Lanyu Shang will present the paper, "A Multimodal Misinformation Detector for COVID-19 Short Videos on TikTok," which she coauthored with PhD student Ziyi Kou, Computer Science PhD student Yang Zhang, and Associate Professor Dong Wang. In the paper, Wang's lab addresses the problem of identifying misleading COVID-19 short videos—such as those on the social media platform TikTok—where misinformation is expressed in the visual, audio, and textual content. To correct this problem, the researchers developed TikTec, a multimodal misinformation detection framework that captures key information from videos and effectively learns (through artificial intelligence) the misinformation that is conveyed by the visual and audio content.
Kou will present the paper, "ExgFair: A Crowdsourcing Data Exchange Approach To Fair Human Face Datasets Augmentation," which he coauthored with Shang, Zhang, Wang, and PhD student Huimin Zeng. According to the researchers, human facial applications are usually biased toward the majority demographic group. To address this limitation, Wang's lab developed ExgFair, a crowdsourcing-based fair data exchange framework, which has been found to not only reduce demographic biases but also improve the accuracy of human facial applications trained on the augmented fair datasets.
The primary research focus of the Social Sensing Lab lies in the emerging area of human-centered AI, big data, and cyber-physical systems in social spaces, where data are collected from human sources or devices on their behalf. The work from the lab addresses the fundamental challenges in social sensing by developing human-centric computing theories, techniques, and systems that reconstruct the correct "state of the world," both physical and social.