Why Vision? Why Research?

In this page, I’d like to share why I decided to (1) pursue graduate school and (2) study vision specifically.

In short, I do vision research because—even after decades of scientific and technological progress—we still don’t have a model that fully explains how the visual cortex works.

There are, of course, many state-of-the-art deep neural networks (DNNs) that can match—or even surpass—human performance on specific visual tasks. These architectures, inspired by the hierarchical organization of the visual cortex, have been remarkably successful in predicting brain activity patterns. Yet, despite these parallels, something essential is still missing. The brain’s visual system remains more flexible, efficient, and robust than any computational model we’ve built. I’m fascinated by this space in-between—the part we haven’t yet captured.

My curiosity began during my freshman year at Johns Hopkins, in Prof. Leyla Isik’s Visual Cognition course, when I first learned about the MIT Summer Vision Project. The idea that scientists in 1966 thought they could replicate the entire human visual system in a single summer—and that we’re still working toward that goal almost sixty years later—captured me instantly. It made me wonder: Why is human vision still so hard to replicate? What makes our brains see the world so effortlessly? Vision, after all, is our main window into the world. To interact with it, we must first perceive light, parse shapes and motion, assign meaning, and use that meaning to act. I’m captivated by how the brain performs this incredibly complex sequence seamlessly—and how it connects to memory, imagination, and perception. Why do we “see” a pyramid when we imagine one? Why do people perceive the same dress as different colors? How do we so vividly recall visual events, or fill in missing details we never actually saw?

These puzzles remind me how much we still don’t know about how seeing turns into understanding. I chose graduate school because I wanted to do more than wonder—I wanted to learn how to build and test models that connect brain computation to perception, and to use those models to understand both human vision and intelligent systems more deeply.