Profs & Pints Northern Virginia: Can Artificial Intelligence “See”?

Profs and Pints Northern Virginia presents: “Can Artificial Intelligence ‘See’?” A look at how humans and artificial intelligence systems interpret the visual world in fundamentally different ways, with Arryn Robbins, assistant professor of psychology at the University of Richmond and cognitive scientist who researches visual attention, perception, and category learning.

[Advance tickets: $13.50 plus sales tax and processing fees. Available at https://events.ticketleap.com/tickets/profsandpints/nv-can-AI-see .]

Artificial intelligence can now identify faces, categorize objects, describe scenes, and outperform humans on certain visual tasks. But does AI actually “see” the world the way that people do? Or does it arrive at correct answers using representations that differ markedly from human perception?

Join Arryn Robbins of the University of Richmond for a fascinating exploration of how humans and AI construct meaning from visual information and a look at comparisons between human perception and AI that reveal just how dynamic and context-dependent our own visual systems really are.

Dr. Robbins, who previously has given excellent Profs and Pints talks on flaws and biases in human visual perception, will draw from research in cognitive science, visual perception, and AI vision systems.

She’ll explain how human perception is not merely a simple recording of the world, but an active process shaped by expectations, context, goals, and recent experience. You’ll learn how humans form flexible mental representations that allow us to recognize objects across changing environments and conditions, and why those representations continuously adapt as we interact with the world.

Many AI systems, by contrast, learn visual categories through statistical patterns in data. They can produce impressive results, but sometimes they also produce strange and unexpected failures, and sometimes they classify images in ways that seem strange to us.

Dr. Robbins will discuss what these differences reveal about the nature of perception itself, and why the mismatch between human and AI representations matters for technologies like self-driving cars, medical imaging, facial recognition, and automated surveillance.

Important for anyone trying to understand the rapidly growing role of AI in daily life, this talk will explore one of the biggest questions in cognitive science and artificial intelligence: What does it actually mean to “see” and understand the world? (Doors: $17, or $15 with a student ID. Listed time is for doors. The talk starts 30 minutes later.)

Image: “Eye Farm” by Nevit Dilmen (Wikimedia Commons).