Bringing Artificial Intelligence Into the Real World

To make science fiction a reality and create AI-powered robots that can move around independently and perform tasks, we first need to solve the problem of computer vision – AI’s ability to recognise the objects that make up our world. 

At the forefront of this rapidly evolving field is Ivan Laptev, Professor of Computer Vision at Mohamed bin Zayed University of Artificial Intelligence in Abu Dhabi (MBZUAI). 

Computer vision in its current form is what enables your phone to search for all the instances of your face in your camera roll. But it is also a big piece of the puzzle when it comes to what’s known as embodied AI, which could markedly transform our everyday lives in the future. As one example, computer vision could make advanced, wearable AI assistants possible. “Imagine a pair of smart glasses that helps you do things you don’t know how to do, like cook a meal to a Michelin standard. Here, the embodied AI can actually watch what you’re doing with your hands and correct you when you make a mistake,” says Ivan. Robots with the physical ability to manipulate the real world – to help you clean up your house, say – are another major potential application. However, their emergence will likely come much further down the line due to the complications that come with developing robots that can manipulate objects with their own hands – or whatever their body parts may be.

But if robots are to be able to manipulate the world at large, they will need to understand it, which is difficult to achieve in a lab setting. What they are exposed to “is nothing compared to the huge and diverse internet data that made generative AI possible,” Ivan explains. Yet it’s these rapid developments in generative AI that could accelerate the robots’ learning. The ability to instantly create lifelike videos could be harnessed to create simulations of any place on earth, a tool that could be used to train robots without ever leaving the lab. While this is a promising solution, generative AI videos still have some way to go. “Video generation is becoming more advanced, but it’s not yet realistic,” says Ivan. “They are not physically consistent – you can turn the camera 90 degrees and turn it back, and see something that wasn’t there before.”

“The field is moving extremely fast, but there are still plenty of interesting problems to solve.” MBZUAI is well-equipped to tackle these challenges, with “around 80 researchers in natural language, robotics, machine learning, and computer vision. If you have a problem, you can always find someone next door to help.” 

As such a young institution – it welcomed its first students in 2020 – MBZUAI is uniquely positioned to define its legacy and shape its global reputation from the ground up. This makes it an ideal setting “for those who want to work hard on high-risk, high-reward projects, and be a part of its growth. The university will make huge leaps over the next 5-10 years.” 

For those who are up to the challenge, MBZUAI offers more than just a booming tech scene. Abu Dhabi has a burgeoning cultural life – it even has its own Louvre, and other museums are in the works. Plus, the region contains nature that is unlike anywhere in Europe. “I’m having picnics in the desert, and then there’s the mountains and the sea,” says Ivan. “You can see dolphins, as well as turtles, who come out at night and lay eggs next to our house.” 

“Before coming here, I knew basically nothing about the region. It was a shock – in a positive way. During the two years I've been here, I’ve never thought of going back. Even in our modern world, you can still discover reality, which is amazing.”