> It even mentions Jupiter's moons later
It is not a coincidence, I reused the example to illustrate my point.
2,040 karma · joined January 4, 2011
> It even mentions Jupiter's moons later
It is not a coincidence, I reused the example to illustrate my point.
we find that there is no distinction between individual high level units and random linear combinations of high level units, according to various methods of unit analysis. It suggests that it is the space, rather than the individual units, that contains the semantic information in the high layers of neural networks.
[1] Intriguing properties of neural networks
If they behave like psychopaths in a systematic way, I don't think they should be granted that right.
Corporations are legally treated like people, that is the point actually [1]. Given that they have similar rights, it is not absurd to expect they also behave like people.
[1] (from wikipedia) A corporation is an organization authorized by the state to act as a single entity (a legal entity recognized by private and public law "born out of statute"; a legal person in legal context) and recognized as such in law for certain purposes.
They do, the list is available here https://www.carsized.com/en/cars/. A quick pipe of this list into wc -l gives you 1311.
In fact, many old school AI, non-ML techniques fit this description. For example, SAT/CSP solving are almost entirely about search, and sometime even about learning, but not in the machine learning sense.
He even mentions "the methods defeated the world champion" at chess as a successful demonstration of AI, but these methods are certainly not what we would call machine learning today (deep search doesn't mean deep learning).
The F-16 encountered months of delays, but the F-35 A/B/C models will, on average, be delayed half a decade.
The F-35 was originally slated to cost $50 million apiece—nearly twice the original cost of the F-16 at today's prices—but the three versions of the plane currently run anywhere from $112 to $120 million each.
Yes, America and her allies have been down this road before, but this time it is a lot rockier.
The fact that you've been reading this comment forever does disprove it since hn has been growing steadily for the past 10 years.
When he says 50%, Neil deGrasse Tyson is quite literally saying "I don't know".
https://en.wikipedia.org/wiki/Crystal#/media/File:Ho-Mg-ZnQu...
This is strangely misleading.
> After 12 further days, the patient was able to reliably increase or decrease his neural activity to hit one of two “target” tones.
Isn't that evidence that this is not random noise?
The only thing that could be left to human interpretation is the "increase of activity" (I'm assuming "activity" is a signal with high dimensionality). But it is easy enough to build some metric, and demonstrate the answers are not random.
I have just enough faith in peer reviewing to hope that the paper would have been rejected without this experiment.
But types are not just useful to specify the content of a variable, they are also useful to specify the required precision.
So, if there is a type for seconds in the type system, should it be a 32 bits int or a 64 bits float? Only the user can say.
I don't think reading cryptocurrency jokes posted on HN is the best way to expose yourself to other cultures.
If by many thousand times bigger, you mean 15 000 times bigger.
I have to admit that it is hard to say if that fits my intuition. But it is also kind of the point: it's much easier to compare heights than volumes.
edit: remove flawed reasoning while I still can :)