Busy GPUs: Sampling and pipelining method speeds up deep learning on large graphs

New technique significantly reduces training and inference time on extensive datasets to keep pace with fast-moving data in finance, social networks, and fraud detection in cryptocurrency.

Lauren Hinkel | MIT-IBM Watson AI Lab • mit
Nov. 29, 2022 ~11 min

A simpler path to better computer vision

New research reveals a scalable technique that uses synthetic data to improve the accuracy of AI models that recognize images.

Adam Zewe | MIT News Office • mit
Nov. 23, 2022 ~7 min


Study: Zebrafish are smarter than we thought

Researchers have discovered that the brains of these simple fish can create three-dimensional maps of their surroundings.

Anne Trafton | MIT News Office • mit
Nov. 18, 2022 ~8 min

Empowering social media users to assess content helps fight misinformation

An experimental platform that puts moderation in the hands of its users shows that people do evaluate posts effectively and share their assessments with others.

Adam Zewe | MIT News Office • mit
Nov. 16, 2022 ~9 min

Solving brain dynamics gives rise to flexible machine-learning models

MIT CSAIL researchers solve a differential equation behind the interaction of two neurons through synapses to unlock a new type of speedy and efficient AI algorithm.

Rachel Gordon | MIT CSAIL • mit
Nov. 15, 2022 ~8 min

Ensuring AI works with the right dose of curiosity

Researchers make headway in solving a longstanding problem of balancing curious “exploration” versus “exploitation” of known pathways in reinforcement learning.

Rachel Gordon | MIT CSAIL • mit
Nov. 10, 2022 ~9 min

In machine learning, synthetic data can offer real performance improvements

Models trained on synthetic data can be more accurate than other models in some cases, which could eliminate some privacy, copyright, and ethical concerns from using real data.

Adam Zewe | MIT News Office • mit
Nov. 3, 2022 ~8 min

Using sound to model the world

This machine-learning system can simulate how a listener would hear a sound from any point in a room.

Adam Zewe | MIT News Office • mit
Nov. 1, 2022 ~8 min


Machine learning facilitates “turbulence tracking” in fusion reactors

A new approach sheds light on the behavior of turbulent structures that can affect the energy generated during fusion reactions, with implications for reactor design.

Adam Zewe | MIT News Office • mit
Nov. 1, 2022 ~8 min

Reprogrammable materials selectively self-assemble

Researchers created a method for magnetically programming materials to make cubes that are very picky about what they connect with, enabling more-scalable self-assembly.

Rachel Gordon | MIT CSAIL • mit
Oct. 20, 2022 ~8 min

/

72