Reasoning skills of large language models are often overestimated
New CSAIL research highlights how LLMs excel in familiar scenarios but struggle in novel ones, questioning their true reasoning abilities versus reliance on memorization.
Department
Electrical Engineering and Computer Science
When to trust an AI model
More accurate uncertainty estimates could help users decide about how and when to use machine-learning models in the real world.
MIT researchers introduce generative AI for databases
This new tool offers an easier way for people to analyze complex tabular data.
Signal processing: How did we get to where we’re going?
In a retrospective talk spanning multiple decades, Professor Al Oppenheim looked back over the birth of digital signal processing and shared his thoughts on the future of the field.
Implantable microphone could lead to fully internal cochlear implants
This tiny, biocompatible sensor may overcome one of the biggest hurdles that prevent the devices from being completely implanted.
The tenured engineers of 2024
Twelve faculty members have been granted tenure in six units across MIT’s School of Engineering.
Study reveals why AI models that analyze medical images can be biased
These models, which can predict a patient’s race, gender, and age, seem to use those traits as shortcuts when making medical diagnoses.
Fotini Christia named director of the Institute for Data, Systems, and Society
Known for building connections between the social sciences, data science, and computation, the political science professor will lead IDSS into its next chapter.
Wireless receiver blocks interference for better mobile device performance
This novel circuit architecture cancels out unwanted signals at the earliest opportunity.
Startup aims to transform the power grid with superconducting transmission lines
VEIR, founded by alumnus Tim Heidel, has developed technology that can move more power over long distances, with the same footprint as traditional lines.
Helping nonexperts build advanced generative AI models
MosaicML, co-founded by an MIT alumnus and a professor, made deep-learning models faster and more efficient. Its acquisition by Databricks broadened that mission.
Arvind, longtime MIT professor and prolific computer scientist, dies at 77
The dedicated teacher and academic leader transformed research in computer architectures, parallel computing, and digital design, enabling faster and more efficient computation.
MIT-Takeda Program wraps up with 16 publications, a patent, and nearly two dozen projects completed
The program focused on AI in health care, drawing on Takeda’s R&D experience in drug development and MIT’s deep expertise in AI.
MIT graduate engineering and business programs ranked highly by U.S. News for 2024-25
Graduate engineering program is No. 1 in the nation; MIT Sloan is No. 5.
Understanding the visual knowledge of language models
LLMs trained primarily on text can generate complex visual concepts through code with self-correction. Researchers used these illustrations to train an image-free computer vision system to recognize real photos.