Members of Associate Professor Dong Wang's research group, the Social Sensing and Intelligence Lab, will present their research at the 2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL 2024), which will be held from June 16-21 in Mexico City, Mexico. NAACL 2024 is the premier venue to present and discuss progress in research, development, standards, and applications of topics in natural language processing and computational linguistics.
PhD student Huimin Zeng will present the paper, "Open-Vocabulary Federated Learning with Multimodal Prototyping." In this paper, Zeng and his collaborators propose a novel AI framework called Fed-MP that can handle open-vocabulary challenges in federated learning, a machine learning technique that allows multiple devices to train a shared model without sharing their data. Fed-MP leverages large multimodal models to understand questions about unseen categories by utilizing a multimodal prototyping mechanism and an adaptive aggregation protocol. This framework allows the global federated model to exploit the semantic knowledge of local data and make better predictions for data from new, unseen categories (i.e., open-vocabulary queries). Fed-MP aims to enhance model generalization and robustness in real-world applications where new and unknown data categories frequently arise.
Informatics PhD student Zhenrui Yue will present the paper, "Evidence-Driven Retrieval Augmented Response Generation for Online Misinformation." In this paper, the researchers introduce a counter-misinformation framework called RARG to provide timely and evidence-based interventions in online discussions. RARG leverages a large database of over one million academic articles to retrieve relevant facts, using them as evidence to generate polite and factual responses through a large language model-based retrieval and response generation process. This framework aims to address the shortcomings of existing methods by incorporating external knowledge to enhance the quality and factuality of the generated responses.
The primary research focus of the Social Sensing and Intelligence Lab lies in the emerging area of human-centered AI, AI for social good, and cyber-physical systems in social spaces. The lab develops interdisciplinary theories, techniques, and tools for fundamentally understanding, modeling, and evaluating human-centered computing and information (HCCI) systems, and for accurately reconstructing the correct "state of the world," both physical and social.