Category: Data & Computing
-
Improving AI models: Automated tool detects silent errors in deep learning training
TrainCheck uses training invariants to find the root cause of hard-to-detect errors before they cause downstream problems, saving time and resources.
-
Off-road autonomy and digital twins: a Q&A with Bogdan Epureanu
Going beyond driving or tele-operating single vehicles, an up-to-date digital environment is needed to help humans operate fleets of autonomous vehicles.
-
Magnetic switch traps quantum information carriers in one dimension
Innovations in quantum sensing and computing could follow the discovery of how chromium sulfide bromide responds to magnetic fields.
-
AI, computation and scientific discovery: A Q&A with Karthik Duraisamy
The director of the Michigan Institute for Computational Discovery and Engineering discusses the institute’s past and future.
-
Boosting AI model size and training speed with lightwave-connected chips
AI growth is capped by data transfer rates between computing chips, but transferring data with light could remove the ceiling.
-
Up to 30% of the power used to train AI is wasted. Here’s how to fix it.
Smarter use of processor speeds saves energy without compromising training speed and performance.
-
Accounting for bias in medical data helps prevent AI from amplifying racial disparity
Some sick Black patients are likely labeled as “healthy” in AI datasets due to inequitable medical testing.
-
Four election vulnerabilities uncovered by a Michigan Engineer
All have solutions, some are implemented.
-
$15M for game theory with AI agents, quantum semiconductors for microelectronics and photonics
The DoD funds efforts to incorporate AI agents into game theory and develop microelectronics that can withstand a hot day on Venus or carry quantum information.
-
AI chips could get a sense of time
Timekeeping in the brain is done with neurons that relax at different rates after receiving a signal; now memristors—hardware analogues of neurons—can do that too.
-
Auto industry deadlines loom for impaired-driver detection tech, U-M offers a low-cost solution
Current technologies already in use could help prevent crashes and deaths linked to impaired driving.
-
Optimized traffic signal timing approach cuts delays 20%, real-world test shows
Communities could reduce costs and cut vehicle emissions—all in the name of shortening your trip.