Antonio Torralba is fascinated by the way humans see and interpret the world. His research interests include computer vision, machine learning, and human visual perception, and he works to create systems that are able to perceive the world in the same manner as humans.
Professor Torralba explains that when machines recognize objects, they are generally performing pattern matching. They memorize a number of appearances that an object might have and match what they see with those appearances, but they don’t understand the visual world as humans do.
On the contrary, humans have visual intelligence. We are capable of understanding objects and environments that we are unfamiliar with, predict what is going to happen next, and make countless other observations about the world. This visual intelligence is something that is still missing in the machines, and is a considerable challenge that Professor Torralba is trying to tackle.
According to Professor Torralba, computer vision is a technology that impacts many different sectors, from health care to security, but, at the moment, autonomous driving has the most powerful influence on the field. “Any company within the sector of autonomy is having a really big impact on the research that is done with computer vision here at CSAIL and all over the world,” Professor Torralba shares.
When Professor Torralba and his colleagues collaborate with industry members, the companies typically approach his team with a specific issue that they want resolved, and his team addresses their needs and uses their techniques to seek answers to their problems. These partnerships are equally rewarding for the CSAIL researchers; Professor Torralba claims that the real-world challenges that industry encounter inspire further research and raise fundamental questions about how vision works.
For Antonio and his team, this means that goals are sometimes defined by a particular obstacle that someone is confronting, and other times, goals arise from fundamental research questions about what they consider critical topics within their area.
Computer vision has seen a significant number of advances in recent years, but there are still many problems that need to be solved. “There are many challenges, and that’s what makes computer vision so exciting,” says Professor Torralba. As computer vision continues to evolve, Professor Torralba’s long term goals are to deepen his understanding of how vision works but also learn more about perception and how these elements integrate with cognition.
Antonio Torralba obtained his Ph.D. in Signal-Image-Speech from Grenoble Institute of Technology in France and has been a member of the MIT faculty since 2005. In addition to his work in CSAIL, he is the Inaugural Director of MIT’s Quest for Intelligence, which aims to discover the underpinnings of human intelligence to build machines that can be utilized for the benefit of society.