Soil and Crop Sciences

Lee Dunnigan

Research focus: Soil plays a vital role in storing carbon, making it essential to understand the processes that influence carbon ecosystem cycling. Traditionally, the remains of plants and animals were thought to be the main drivers of soil carbon dynamics. However, new research suggests that microbial byproducts may have an even greater impact on carbon cycling. Studying these compounds is challenging due to the complexity of soil, but this research aims to improve chemical analysis and machine learning techniques to better identify and classify them. By doing so, it will provide deeper insights into how soil stores carbon over time to reduce global atmospheric carbon. 

My future goals: My goal is to earn my PhD and pursue a career at the intersection of environmental science and machine learning. I’m particularly interested in applying data-driven approaches to improve our understanding of complex environmental systems and fundamental biogeochemistry at the molecular level. I ultimately hope to become an expert who can help address pressing climate and sustainability challenges.

Why AI4S NRT: The NRT program caught my interest because of its clear goal of building a community of grad students and faculty with shared interests in AI and sustainability. When I was encouraged to check it out, it immediately felt like the perfect fit. It is a rare chance to get connect with many members of Cornell whose research isn’t directly related to mine but who still offer incredible guidance, knowledge, and genuine support for the success of my work.

Computer Science

Christian Belardi

Research focus: My research explores a variety of topics in machine learning, particularly generative models, as well as how they can be applied to solve complex challenges in scientific domains. Though I work on a number of problems, my work enabled by the NRT fellowship focuses on developing approaches that will enhance the existing capabilities of microscopists. While capabilities for text and image domains continue to develop rapidly, models and methods for specialized scientific problems remain underdeveloped. Our objective is to fill this gap for Multislice Electron Ptychography (MEP). Our current work greatly improves atomic resolution reconstructions of crystal structures from MEP diffraction patterns. 

My future goals: My primary goal is to contribute to the field of machine learning through fundamental research. Beyond this, I hope to help address humanity’s critical challenges by accelerating scientific discovery with machine learning and, more generally, computational methods.

Why AI4S NRT: The NSF NRT fellowship presents a unique opportunity to interact and collaborate with researchers beyond my own field. Our work on MEP is truly an interdisciplinary collaboration, made possible by unique programs like Cornell’s AI for Sustainability initiative. I’m grateful to be a part of this community developing innovative solutions for global sustainability challenges.

Computer Science

Joshua Fan

Research focus: My research focuses on making AI work on problems for which little data is available. Recent advances in AI have relied on massive datasets, but for many important problems in sustainability, we may only have a few hundred observations or human annotations. To overcome data shortages, I develop methods to embed pre-existing scientific knowledge into our models, make models simpler and more interpretable, and leverage unlabeled data such as satellite images. My methods have been used to simulate how carbon is sequestered in soils, map fish farms from satellite images, and forecast crop yields given climate and soil data. 

My future goals: My goal is to pursue high-impact research at the intersection of AI and sustainability sciences (e.g. agriculture, ecology, climate science). Ideally, I would like to work on problems with high societal impact, such as preventing food insecurity, using satellite images to assist with humanitarian relief efforts, or predicting extreme weather events. I would also like to make advancements in fundamental AI research, with the aim of making AI models more interpretable and easier-to-use for scientists and policymakers. While these are ambitious goals, I enjoy working with interdisciplinary teams to tackle these challenges.

Why AI4S NRT: I am very excited by the depth of interdisciplinary training and collaboration that the NRT program fosters. In order to solve challenging problems beyond the scope of traditional AI methods, collaborations with domain experts in soil science, ecology, and agricultural economics have been particularly crucial to generate solutions that are informed by scientific insights. The NRT program is a great opportunity for me to extend these collaborations and gain more interdisciplinary knowledge on AI and sustainability.

Systems Engineering

Guoqing Hu

Research focus: My research focuses on using model predictive control and machine learning techniques to optimize energy efficiency and robust control systems in critical applications. Specifically, I have focused on applying MPC to develop the optimal control decision for sustainable energy management in semi-closed greenhouses, ensuring efficient crop production while minimizing resource usage. Additionally, I have worked on creating advanced multi-zone building control systems that address thermal comfort while handling operational uncertainties through data-driven robust control methods. Lastly, I have integrated AI frameworks with MPC to enhance energy-efficient food production in built environments, advancing sustainability and automation in controlled settings. 

Research focus: My research focuses on using model predictive control and machine learning techniques to optimize energy efficiency and robust control systems in critical applications. Specifically, I have focused on applying MPC to develop the optimal control decision for sustainable energy management in semi-closed greenhouses, ensuring efficient crop production while minimizing resource usage. Additionally, I have worked on creating advanced multi-zone building control systems that address thermal comfort while handling operational uncertainties through data-driven robust control methods. Lastly, I have integrated AI frameworks with MPC to enhance energy-efficient food production in built environments, advancing sustainability and automation in controlled settings. 

My future goals: In my future career, I aspire to transition into the industry to extend the practical applications of MPC in modern greenhouses, making them more efficient and scalable. My goal is to test the feasibility of crop models by utilizing data collected from local greenhouses and evaluating the complexity of integrating advanced control systems into these environments. Additionally, I aim to build a comprehensive crop library by collecting diverse crop data from varied regions, enabling the development of universal models that can be applied globally. These efforts would drive innovation in sustainable agriculture and advanced greenhouse automation.

Why AI4S NRT: I am choosing the NRT fellowship to advance sustainable agriculture through innovation in greenhouse automation and crop management. My research focuses on using MPC and machine learning to optimize greenhouse operations, improve energy efficiency, and address resource challenges in modern agriculture. With support from NRT, I aim to test data-driven crop models using local greenhouse data, evaluate the integration of control systems, and develop a global crop library for universal application. This fellowship will provide the resources and interdisciplinary collaboration necessary to translate my research into practical solutions, driving advancements in agricultural development and sustainability.