Image
alt="MIT CSAIL researchers helped design a new technique that can guarantee the stability of robots controlled by neural networks. This development could eventually lead to safer autonomous vehicles and industrial robots (Credits: Alex Shipps/MIT CSAIL)."
CSAIL article

Neural networks have made a seismic impact on how engineers design controllers for robots, catalyzing more adaptive and efficient machines. Still, these brain-like machine-learning systems are a double-edged sword: Their complexity makes them powerful, but it also makes it difficult to guarantee that a robot powered by a neural network will safely accomplish its task.

Image
The Grasping Neural Process uses limited interaction data to help robots understand unclear objects in real-time (Credits: Alex Shipps/MIT CSAIL).
CSAIL article

When robots come across unfamiliar objects, they struggle to account for a simple truth: Appearances aren’t everything. They may attempt to grasp a block, only to find out it’s a literal piece of cake. The misleading appearance of that object could lead the robot to miscalculate physical properties like the object’s weight and center of mass, using the wrong grasp and applying more force than needed.

Image
Say More Logo
External articles

Daniela Rus’s dream is to imbue the power of robotics with the wisdom of humanity. She runs MIT’s Computer Science and Artificial Intelligence Laboratory. As part of his ongoing series on the promise and perils of AI, Globe Ideas Editor Brian Bergstein talks to Rus about her new book “The Heart and the Chip.” She says robots won’t just do our chores and work in our factories; they can teach us how to hit tennis balls like Serena Williams and defy gravity like Iron Man. She says your car won’t just drive you around — it might also be a friend. 

Image
alt="MIT researchers’ "consensus game" is a game-theoretic approach for language model decoding. The equilibrium-ranking algorithm harmonizes generative and discriminative querying to enhance prediction accuracy across various tasks, outperforming larger models and demonstrating the potential of game theory in improving language model consistency and truthfulness (Credits: Alex Shipps/MIT CSAIL)."
CSAIL article

Imagine you and a friend are playing a game where your goal is to communicate secret messages to each other using only cryptic sentences. Your friend's job is to guess the secret message behind your sentences. Sometimes, you give clues directly, and other times, your friend has to guess the message by asking yes-or-no questions about the clues you've given. The challenge is that both of you want to make sure you're understanding each other correctly and agreeing on the secret message.

Image
Three new frameworks from MIT CSAIL reveal how natural language can provide important context for language models that perform coding, AI planning, and robotics tasks (Credit: Alex Shipps/MIT CSAIL, with components from the researchers and Pixabay).
CSAIL article

Large language models (LLMs) are becoming increasingly useful for programming and robotics tasks, but for more complicated reasoning problems, the gap between these systems and humans looms large. Without the ability to learn new concepts like humans do, these systems fail to form good abstractions — essentially, high-level representations of complex concepts that skip less-important details — and thus sputter when asked to do more sophisticated tasks.