In our already metricized world, we’re familiar with complex aspects of our experience being made commensurable: that is, turned into numbers for comparison and ranking, whether that’s in a national league table of school performance or in …real challenge is interpreting these millions of weights in a way that is accessible to human reasoning. The network can’t tell us why a particular pattern in any layer is significant: it delivers a prediction, not an explanation. So, while…
AlexNet
fictional · 2 mentions across 1 reading
In this course
AlexNet refers to the deep convolutional neural network that won the 2012 ImageNet competition and catalyzed the modern deep learning revolution in computer vision. The course readings invoke it to illustrate the interpretability problem central to AI systems—that while networks like AlexNet can produce accurate predictions from massive datasets, they remain opaque black boxes unable to explain their reasoning to humans. This gap between performance and explicability becomes crucial when such systems are deployed in high-stakes contexts where decisions affecting society require not just accuracy but legible justification.
Mentioned in 1 reading
Appears alongside
People mentioned in the same passages — sorted by co-occurrence weight.