neural nets
Computational models loosely inspired by the brain, central to modern machine learning.
As defined in Michael Nielsen
Where it comes up · 4
- 84 - Landscapes Of Mind2 claims
- Michael Nielsen1 claim
- Kevin Kelly on Advice, AI, and Technology1 claim
- 162 - Medical Intelligence1 claim
Most important claims · 5 of 5Search all claims for “neural nets” →
- Deep learning neural nets are only as good as their inputs, but people don't own, possess, or control their own medical data, which is scattered with no single home; consolidating all of a person's data under their ownership would make algorithmic outputs far more ideal and less compromised, enabling a virtual medical coach to prevent illness.162 - Medical Intelligence
- Current AI relies on neural nets needing billions of data points, unlike a toddler who learns cat-versus-dog from a dozen examples; once few-shot training arrives, training sets will be tiny and artists will clamor to be included to be counted among the few influential ones—making intellectual-property objections a distraction.Kevin Kelly on Advice, AI, and Technology
- Modern neural-net AI is largely impenetrable to us—it makes decisions that affect us without our being able to understand how they are made—which is why DARPA funds 'explainable AI' that stacks another AI layer to extract intermediate reasons, a self-monitoring loop that is almost the beginning of a kind of consciousness.84 - Landscapes Of Mind
- Putting his neural nets textbook online for free generated far greater impact and ultimately benefited Nielsen more financially — through reputation, opportunities, and jobs — than any amount of royalties would have, so for authors paid royalties make little sense even though publishers' incentives differ.Michael Nielsen
- A black-box machine-learning system used to decide prison parole by minimizing recidivism was found to be using race as a relevant variable, becoming in effect a racist AI, and because such systems are not inspectable we generally cannot know what variables they rely on—illustrating that AI's unknowability is a serious concern even short of any singularity.84 - Landscapes Of Mind