Machine learning models are easy to trick.
“But adversarial examples keep cropping up: seemingly trivial changes to objects (even a single pixel) can stump an AI or send a self-driving car barelling [sic] through a stop sign.” | learn more
“But adversarial examples keep cropping up: seemingly trivial changes to objects (even a single pixel) can stump an AI or send a self-driving car barelling [sic] through a stop sign.” | learn more
This is the most recent issue of the most in-depth newsletter about AI that I’ve ever seen. The author Nathan Benaich deserves a standing ovation. | learn more
I can’t explain this, you’ll just have to try it for yourself. | learn more
The IEEE is keeping score of the battle between Doctors and Artificial Intelligence. The results are interesting, and the best part is we all win! | learn more
“Three years after IBM began selling Watson for Oncology to recommend cancer treatments, it’s falling short of the lofty expectations IBM created for it.” After reading this, I feel like the criticisms are valid, but mostly a mismatch of marketing […]
German AI startup TwentyBN has created an AI algorithm that can identify what’s happening in a video stream in real time. An example: “Trying to pour water into a cup, but missing so it spills next to it.” They claim […]
I’ll admit it – I’m guilty of being over-enthusiastic about AI ever since reading Nick Bostrom’s Superintelligence a few years ago. Robotics engineer and AI smarty-pants Rodney Brooks published this post in the MIT Technology Review that pours some much-needed cold water […]
These days, SoftBank’s outsized investments in tech companies regularly make the news. The uniting concept behind them all is less apparent, however. This NYT article helps put them in context – the plan is to get ahead of, and profit […]
The 180 year old company had trouble getting AI talent to Iowa, so they set up shop in the valley instead. They are focusing on technology for precision agriculture, managing crops at the plant level for increased yields and lower costs. |learn […]
With a training dataset of only 67 MRI scans, “The AI was able to diagnose Alzheimer’s 86 percent of the time. More importantly, it was able to detect mild cognitive impairment 84 percent of the time, making it a potentially […]