Chen-yi Lu

AI takes a lot of work to make it work. Currently, you have to throw a ton of resources — time, people, computing power, energy utilization — at training, supervising, and optimizing your models to get AI platforms and apps that can perform intelligently in the real world. Chen-yi Lu, a second-year PhD student and graduate research assistant in Purdue Engineering’s Department of Agricultural and Biological Engineering, wants to make AI models that can work more on their own accord.

What is your research focus? What investigative avenues are you pursuing?

My research is focused on data-efficient learning algorithms and deep representation learning for computer vision tasks, including image semantic segmentation, video editing, and image generation. Furthermore, I am interested in research on multimodal learning, specifically text-to-image diffusion models and text-guided video editing. I hope to develop robust AI models for computer vision tasks that do not depend on extensive supervised data.

What spurred your interest in this topic, and line of research?

I led a project focused on developing a small device for greenhouse environments capable of detecting and monitoring insect pests using AI models. During this project, I dedicated significant time to annotating each insect category and their location. This experience led me to question the feasibility of developing AI models that are less reliant on fine-grained supervised data. After all, humans can infer information relatively quickly without needing to reference thousands of images of the same class of insect; why can't machines do this as well?

Why pick Purdue to continue your studies as a graduate student?

In addition to Purdue's rankings and academic resources, I selected this university due to its notable school spirit, which emphasizes being hardworking, cooperative, and supportive not only in academics but also in athletic events and overall well-being.

When did you first get interested in engineering and science?

A fifth-grade project, in which we were guided by our teacher to construct a water rocket, was a formative experience. The process of continually refining and adjusting the design to achieve greater flight altitude instilled a deep fascination with engineering and science, which continues to motivate me today.

What’s it like studying at Purdue?

The academic environment here is incredibly stimulating. It's inspiring to be surrounded by such talented professors and peers who are not only driven to achieve academic success but also share a genuine curiosity and passion for solving problems and gaining a deeper understanding of the world.

Beyond subject matter and field, what else have you learned at Purdue?

The seminars and on-campus guest lectures have broadened my understanding of both my research field and the focus of other disciplines.

What is the Purdue research environment like?

I gained experience working with the Purdue patent office, where they patented my published works. The patent office staff were highly professional and adept at rephrasing the language in my research articles into patentable formats. Following my engineering internship as a research scientist with Adobe, my lab was awarded a gift fund. I was also fortunate enough to travel to Milan to attend a conference on computer vision where I published my work last year.

What advice might you give to other students deciding where to attend graduate school?

The most important factor when deciding where to attend graduate school is selecting an advisor, who will be an essential mentor throughout your academic journey. Purdue University has an excellent faculty with diverse and extensive research interests, making it an ideal institution to begin a graduate student career.

What about the future? What are your goals; what are you looking to accomplish?

I aspire to continue my research in deep representation learning, either in an industry setting or within academia, where I would have the opportunity to teach and guide motivated students.

Might you share with us a little window into your personality: some distinctive trait, habit of mind, hobby/pursuit outside work…?

When I'm not working or researching, I like to travel, hike, rock climb, and spend time with friends. I'm also an avid reader of fiction and philosophy.