Zaiwen of Huazhong Agricultural University (HZAU), China

Abstract: Integrating deep learning technologies into agriculture promises to pave the way for a more comprehensive form of agricultural intelligence—capable of processing diverse inputs, making decisions, and potentially overseeing entire farming systems autonomously. In particular, Large Language Models (LLMs) have introduced a new paradigm in smart agriculture. These models, built on the Transformer architecture, incorporate a vast number of parameters and have undergone extensive training, often demonstrating excellent performance and adaptability. This makes them effective in addressing agricultural challenges, especially where data is limited.
In this talk, the challenges and key tasks of applying LLMs in agriculture are first analyzed. The framework of the ShizishanGPT, developed by Huazhong Agricultural University, is introduced. Then, several core technologies within ShizishanGPT are presented, including data governance, construction of knowledge graph, tool learning, and AI agents. Subsequently, a couple of key scenarios that the ShizishanGPT are employed in the digital production of large fields will be introduced -- for example, precision crop breeding, crop modeling, yield prediction and the identification of crop diseases and pests. Lastly, several scenarios that the ShizishanGPT are used in the intelligent greenhouse will be introduced, such as, agricultural inspection robots and environmental control system.

Biography:

Zaiwen Feng is an Associate Professor at the College of Informatics, Huazhong Agricultural University (HZAU). Before joining HZAU in 2020, he worked as a Lecturer at the State Key Laboratory of Software Engineering (SKLSE), Wuhan University, China, from 2009 to 2016, and as a Research Fellow in the STEM unit at the University of South Australia (UniSA) from 2017 to 2019. He also served as a Postdoctoral Visiting Scholar at the School of Information Systems, Queensland University of Technology (QUT) in 2014. He received his ME and PhD degrees in Computer Science from Peking University in 2006 and Wuhan University in 2009, respectively. His current research interests include knowledge fusion, Retrieval-Augmented Generation with Large Language Models, causal inference, and their applications in Smart Agriculture.