"We show that backpropagated neural networks trained on a variety of datasets - which could be
disjoint and unrelated - diverse hyper-parameter settings, initializations and regularization methods, often learn an architecture-specific, layer-wise similar, low-rank joint subspaces (we refer to this as the Universal Subspace). We provide the first large-scale empirical analysis - across a diverse set of models - that neural networks tend to converge to these joint subspaces, largely independent of their initialization or the specific data used for training."
I think you have to read this one generously. The claim isn't about how COBOL works, except incidentally. It's more along the following lines:
COBOL doesn't have a default date/time type
As such implementation decisions are left to the implementor
The implementors* of the SS system chose 1875 as the epoch date for reasons
*I made a lot of money in 1999. The original implementors of SS probably used something else ("it'll be rewritten before this is a problem" was essentially the whole raison de etre of Y2K). The 1875 thing, if it's a thing, was probably the result of Y2K work. But I have no direct knowledge of these matters.
Only replying because playing same thought in head. Once two points relation exists. Simplest form distance. Either may be considered centre. Infinite points exist along that relation. And more can be imagined. The set of all points having the same distance from A as the distance from A to B etc.
The whole is better than the sum of the parts. You have nostradamus which predicts the future, fork bot which plays all playable presents, and the psychic friends network which rewrites the past. All under the watchful eye of the matrix.
There's something beautiful here and you honestly couldn't make it up.
“Future users of large data banks must be
protected from having to know how the data is
organized in the machine (the internal
representation)… Activities of users at
terminals and most application programs
should remain unaffected when the internal
representation of data is changed and even
when some aspects of the external
representation are changed…”
Roll the related data up into materialized views for read performance.
But this is a bit silly surely? If they are different ways of talking about the same thing then one might as well label oneself a 'mentalist' and say "physical processes don't 'have mental effects', they just flat out are mental processes."
I think what you're going for is a species of monism maybe? The idea that Thought and Extension are attributes of one Substance?
"I am no judge of poetry – the only poem I ever wrote was produced when I was body and soul in the gilded harness of Dame Laudanum – but I think Mr Kavanagh is on the right track here. Perhaps the Irish Times, timeless champion of our peasantry, will oblige us with a series in this strain covering such rural complexities as inflamed goat-udders, warble-pocked shorthorn, contagious abortion, non-ovoid oviducts and nervous disorders among the gentlemen who pay the rent"
I'd add a note to say some more or whisper Latin at the door
but I have promises to keep and Myles to go before I sleep...
Let's say for a second the King (or Crown, it makes no odds) owns your swans, but you'd rather he not own them. What'd be your recourse?
Presumably the King (or Crown, it makes no odds) once owned all the swans in Ireland too. Then one day didn't. Now those swans are free swans living in an actual, honest to God, Republic. Technically the ones in the North are still his but their ownership status is disputed. The swans that fly back and forth I'm not sure what to do about, post Brexit arrangements being shoddy at best.
All this talk of ownership seems to confuse power with right, right with tradition, tradition with law, law with power.