Personalization features can make LLMs more agreeable
The context of long-term conversations can cause an LLM to begin mirroring the user’s viewpoints, possibly reducing accuracy or creating a virtual echo-chamber.
IDSS
3 Questions: Using AI to help Olympic skaters land a quint
MIT Sports Lab researchers are applying AI technologies to help figure skaters improve. They also have thoughts on whether five-rotation jumps are humanly possible.
Study: Platforms that rank the latest LLMs can be unreliable
Removing just a tiny fraction of the crowdsourced data that informs online ranking platforms can significantly change the results.
New method improves the reliability of statistical estimations
The technique can help scientists in economics, public health, and other fields understand whether to trust the results of their experiments.
A smarter way for large language models to think about hard problems
This new technique enables LLMs to dynamically adjust the amount of computation they use for reasoning, based on the difficulty of the question.
New control system teaches soft robots the art of staying safe
MIT CSAIL and LIDS researchers developed a mathematically grounded system that lets soft robots deform, adapt, and interact with people and objects, without violating safety limits.
Creating AI that matters
How the MIT-IBM Watson AI Lab is shaping AI-sociotechnical systems for the future.
3 Questions: On biology and medicine’s “data revolution”
Professor Caroline Uhler discusses her work at the Schmidt Center, thorny problems in math, and the ongoing quest to understand some of the most complex interactions in biology.
Simpler models can outperform deep learning at climate prediction
New research shows the natural variability in climate data can cause AI models to struggle at predicting local temperature and rainfall.
A new way to test how well AI systems classify text
As large language models increasingly dominate our everyday lives, new systems for checking their reliability are more important than ever.