Artificial intelligence is in our phones, beginning to control our money, and giving robots some decision-making powers. But Carnegie Mellon professor Manuela Veloso, incoming head of JPMorgan’s AI research division, is frustrated that things aren’t further along. “I came from the Boston airport last night and I didn’t see a single mobile robot anywhere on my way here,” she said.
After her decades of work in AI research, Veloso is ready to see more mobile robots in our everyday lives. She envisions a future where everything with wheels, from suitcases to shopping carts, will automatically follow you at a single command. “Every time I enter a supermarket and I push this cart, I say, ‘Why can’t this cart follow me?’” said Veloso. “These carts should all be automated.”
She is focused on turning her complaints into solutions. Veloso is working on making smart mobile robots more accessible and autonomous. If you visit her at Carnegie Mellon, a robot will greet you and lead you to her office. It might even grab you coffee.
In order to improve their autonomy, Veloso trained her robots to let you know when they don’t know something, and ask for help. “AI might not be able to do the whole task, but if you enable the ability to ask for help or plug in other things, then it is able to do the whole [task],” she said.
For example, her office robots don’t have arms, but they are programmed to enter the kitchen and ask any human within earshot to put a cup of coffee in its basket. By recruiting some help, it can complete its task of fetching coffee and autonomously navigating to its final destination.
Want to hear more from Manuela Veloso? We hosted a live Ask Me Anything with her on Facebook today. You can watch the full video here.
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
The therapists using AI to make therapy better
Researchers are learning more about how therapy works by examining the language therapists use with clients. It could lead to more people getting better, and staying better.
DeepMind says its new language model can beat others 25 times its size
RETRO uses an external memory to look up passages of text on the fly, avoiding some of the costs of training a vast neural network
Meta’s new learning algorithm can teach AI to multi-task
The single technique for teaching neural networks multiple skills is a step towards general-purpose AI.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.