An autonomous-vehicle system that directly explains its decisions in a way that can be interpreted by humans could improve driving safety.
Category: robotics/AI – Page 39
Uber and Wayve to Start Testing Robotaxis in London
The city’s medieval layout is a dense maze filled with cars, buses, delivery scooters, bikers and jaywalking pedestrians. There are roundabouts upon roundabouts. Restricted bus lanes emerge and disappear again. On one-lane streets, drivers traveling in opposite directions intersect and must coordinate in an unspoken language to get through.
The failure rate for a driver’s license test in London is just over 50 percent.
This week, a small number of autonomous robotaxis will enter the fray, carrying London passengers for the first time. Fifteen cars operated by Uber and the British autonomous vehicle start-up Wayve will be available for ride-hailing through Uber’s app as part of a safety testing period. The cars will have a safety driver present to take over if needed, as government regulators weigh eventually approving wider deployments.
Did the Unabomber Predict the Singularity? Ted Kaczynski’s Warning
16, he entered Harvard. At 25, he was the youngest assistant professor Berkeley had ever hired. At 27, he walked away from all of it into a Montana cabin with no electricity and no running water.
Then he started mailing bombs.
Ted Kaczynski killed three people and injured twenty-three others. Nothing excuses that, and this is not an attempt to.
But in 1995 he traded an end to the killing for the publication of his manifesto, and there are two passages in it I have not been able to shake. Not because they are unhinged. Because of when they were written and what they describe.
I am deliberately not quoting them here. You should read them cold, with no name attached, and work out for yourself what you are looking at.
Then the uncomfortable question arrives on its own. Was the Unabomber a murderous neo-Luddite whose ideas deserve to be buried with him, or was he among the first to see the #Singularity coming and recoil at it?
My answer will not satisfy either camp.
Shining Science (@ShiningScience) on X
Shining Science on X: 🚨 A groundbreaking MIT brain-scan study shows relying on ChatGPT weakens critical thinking and memory wiring.
A recent MIT study, titled The Cognitive Cost of Using LLMs, monitored the brain activity of 54 students using electroencephalography (EEG) devices.
Researchers from the MIT Media Lab discovered that participants who consistently relied on ChatGPT for essay-writing tasks demonstrated significantly lower neural engagement in brain regions tied to memory and analytical reasoning.
A new kind of AI that does its thinking cheaply without words
There may soon be a new kind of artificial intelligence in town, one that uses a novel approach to thinking that could save massive amounts of computing power and money.
The AI that most people use every day, like ChatGPT, Claude or Google Gemini, relies on large language models that can work through difficult problems step by step. While this may ultimately give us the answers we are searching for, the process can increase response times and use a lot of computing power.
So scientists at AI company Pathway decided to test an alternative approach. The research and technical foundations of the team’s work are in a paper posted on the arXiv preprint server.
AI-informed AdaptiveFlow redefines large-scale cloud computing for drug discovery
Researchers today announced AdaptiveFlow, an AI-informed platform that can virtually screen billions of drug-like molecules with a 1,000-fold reduction in computational costs over existing methods. Developed and validated by scientists from St. Jude Children’s Research Hospital, University of Pavia, Dana Farber Cancer Institute and Harvard Medical School, AdaptiveFlow allows prohibitively expensive ultra-large virtual drug screens to be conducted routinely. The open-source platform was published today in Nature Biotechnology.
The platform’s framework demonstrated linear scaling up to 5.6 million virtual central processing units (CPUs)—a new benchmark for cloud-based drug discovery—allowing billions of molecules to be screened without loss of efficiency. As proof of concept, the team identified potent inhibitors for existing and emerging cancer targets for which few inhibitors are known.
“AdaptiveFlow is the next generation in automated drug discovery platforms for routine ultra-large virtual screenings,” said co-corresponding author Christoph Gorgulla, Ph.D., Center of Excellence for Data-Driven Discovery, St. Jude Department of Structural Biology. “With this platform, we are able to screen 69 billion molecules, representing the largest ready-to-dock library in the world.”
Human model of childhood dementia from drug screen and AI
In a new study published in the prestigious Nature Communications journal, medical researchers used patient-derived brain cells grown in the lab, combined with advanced imaging and artificial intelligence, to rapidly test approved drugs and pinpoint those that improve brain cell health.
The research focused on Sanfilippo syndrome, a devastating childhood dementia that causes progressive loss of memory, behavior and physical abilities. There are currently no widely available treatments.
Sanfilippo syndrome is one of more than 100 genetic disorders that together affect 1 in 2900 children in Australia and half of all children with dementia die by the age of ten.