On June 18, the seventh seminar of the "AI for Ecology and Environment" Academic Week, part of the "Four Decades of Environmental Science, Shaping the Future Together" series, was held in Room 209 of the School of Environment (SOE). This event, centered on the theme of "Data-Driven Environmental Management", featured a compelling presentation by Professor Bi Jun from Nanjing University, and was chaired by Professor Xu Ming, SOE Associate Dean and Director of Center of AI for Ecology and Environment at Tsinghua University. Nearly a hundred faculty members and students attended the event.

Seminar br Bi Jun

Bi Jun highlighted the critical data needs and challenges in environmental management, providing a detailed examination of how data is applied and valued in the field. He pointed out that achieving the goals of the Beautiful China initiative is inextricably linked to data-driven management, emphasizing the pressing need for the digitalization of environmental practices. Data, he emphasized, is the key to balancing supply and demand and serves as a vital link between the implementation of advanced technologies and the demands of effective environmental governance. Currently, environmental management faces three major data-related challenges: data gaps, data overload, and data distortion.

To tackle the problem of data gaps, Bi Jun used real-world examples to demonstrate how machine learning and big data analysis can be employed to complete carbon emission data, thereby improving both coverage and reliability.

To tackle the issue of data overload, Bi Jun explained how key information can be extracted from the vast amounts of social media data to improve the scientific management of climate change risks. He demonstrated the use of natural language processing and sentiment analysis models to identify core topics and emotional trends from millions of climate-related Weibo posts. This approach aids in identifying opinion leaders, analyzing their rhetoric, and developing effective public awareness strategies for climate change.

To address the challenge of data distortion, Bi Jun utilized unsupervised learning and anomaly detection techniques to identify and correct fraudulent reports of hazardous waste, thereby enhancing regulation efficiency. He demonstrated the effectiveness and precision of the machine learning-based "Risk Alert System for Fraudulent Declaration of Hazardous Waste by Enterprises" in real-world applications, significantly advancing the intelligence level of environmental monitoring.

In his conclusion remarks, Bi Jun outlined the promising future applications of artificial intelligence and big data in environmental management. He highlighted that the value of data in this field hinges on its breadth, accuracy, and depth. He emphasized that effective data governance is crucial for unlocking data's potential, which in turn can significantly enhance both the efficiency and effectiveness of environmental management.

Following the report, Bi Jun engaged in a lively discussion with faculty and students on the application of artificial intelligence in environmental data and the prospects for interdisciplinary development. Xu Ming remarked that digitalization and intelligence are key directions for future ecological and environmental governance. He stated that the School of Environment will continue to foster academic exchanges and collaborations, contributing wisdom and strength to build a Beautiful China.

The attendees