Blog

Neural data analysis; where is it?

DSCN9545nw
[Koolhaas locator question]

As [Big | Smart | Process | Socmed] Data Analysis spreads, like this:
age-of-miss-america_murders-using-a-handgun I wondered where the neural network angle was, or is.

Remember, two decades back, at universities there was so much clamour about Neural Networks and their potential, and how credit card companies were the flagship applications with their pattern recognition work ..?
All that had never grown to full fruition, to full potential – as proven by the lack of ubiquitous visibility in our daily lives; if neural networks had been so successful, we’d had seen it on and in just about any IoT device already today.

But we don’t.
Hence, neural networks as a community has missed the boat, somewhere. Would anyone have a good pointer to the algorithms behind the said [Big | Smart | Process | Socmed] Data Analysis, and/or which actual algorithms are behind it ..? Either statistical non-linear regressions of sorts (including outlier detection and semi-smart pattern recognition), or neural networks, or …? Still some non- or true visual Kohonen networks out there ..?
I really would want to hear from you.

Oh and while you’re at it, would you also have good pointers to freeware tools to tinker with the algorithms ..?

Donuts for capital requirements

20130211_144900[1]
[Slippery world, trying to balance on stelts, as long as that goes]

… Suddenly, I realized that the approach towards balanced capital requirements calculations through the combination of four quadrants of external vs. internal, data-driven versus ‘fuzzy’ scenario analysis, is flawed when depicted as a blot around the centre of the said four quadrants or axes.

What is needed, is in fact a donut. As in:
DonutII
Where the ideal is a combination indeed of all sides of the picture, to a quality and coverage as far out as possibly one can get, but one wouldn’t want to lose all that is gained by getting away from the centre when reverting to it to combine the conclusions. One will want to have a donut, capturing all the peculiarities and ground covered. Into one multi-multi-faceted capital requirements calculation.

If that leads to an overall picture so fuzzy and risky that no-one would be able to conclude anything from it … then you’ll have succeeded.
In capturing the future, to a level of reliability that’s still very close to, but off from zero.
The fuzzier it gets, the more demonstration you get it, in commensurate degree.

That would be an achievement!
Hence as a token of appreciation, get yourself a donut!

Mgt > control

Is this management, or control?
DSCN2059
[Hospital St. Pau, Barça]

Just a note, on errors (not minor omissions) that I see over and over again:

‘Management’ ≠ ‘Control’

Rather, Management >> Control. Control is not allowing exceptions or deviations of predetermined values beyond a predetermined bandwidth. Management is dealing with whatever value comes out of your measuring system, whether one can steer the inputs to and operations of the production system so the outcomes come back into the fold or not.

Control, one does out of fear. The fear of not being in control (sic) over one’s own destiny, as if that is completely tied up with the workings of the ‘managed’ system, as if systems Compliance were the ultimate you can achieve in life. If so, one should discuss suicide; that shows utter control as well as categorical prevention of any future mishap to one’s plans. Control is about the fear of not being able to deal with deviations for which there is no rule, i.e., for which one as a drone following set detailed orders only (less than a robot, like an inhumane not really thinking let alone ethically operating machine); showing one’s incapacity for dealing with the real world. Admitting one would be happier as the subject in the Truman Show. Not wanting to escape.
Management, as above, is about being capable of understanding the system(s), and being able to see the innate instable nature of life (as a system, too), and dealing with all that comes at you through insight. Having ‘control’ and the rules for the set pieces, but having that, too, for the petty little uninteresting parts of life. The rest, is an adventure. Bring it on!
So, management is so much more than just mere simple control. If one would look down on those that desperately vie for control over … in their ‘management’ positions, a wry smile will do. Avoid them. Spiritual death by crushing bureaucracy is what they spread.

Which, by the way (?), clarifies also our message, in many a previous post, that ‘risk management’ gets to be more and more what ‘first line’ ‘management’ is about in these times of knowledge workers / professionals. The latter, know very well what to do and how to do it; probably better than you (because you apparently were/are more valuable not doing it yourself…). Hence, they require you to defend them against the outside world, starting with the other departments around them. And, your job as manager is to just grind off the rough edges of what your teammates (sic) do as the knowledge workers are managing their work not controlling it – the uninteresting controllable part they leave to machines (and indistinguishable AI ever more).
Facilitate, grind off the rough edges, and defend against ‘asteroid’ impacts from all around. [Actual decisions, cutting through dilemmas (if none, than it’s mere logic), is Leadership work, not for shop-attending managers.]
The latter, of course, is risk management. NOT to be done in separate ‘second line’ departments; if so, they’ll be useless overhead burdens (and much cost!), only to be facilitated, coordinated there – when not if first-line managers think they can do it better via their own methodology and practices, they are right…
Borne out, after writing this, by Seth’s blog again.

So, since they started, are hospitals about management or control ..? Well, control what they can, i.e., the taking away the negative impacts of outside air as at St. Pau by bringing the hospital under ground (yes in the picture, it’s the roof that appears to be street level ..!) and also enabling fresh air to be enjoyed by those who’d want to wander around on-site, both functions in one place not spread out. But not controlling (here in the picture!!) the health of their patients, managing that. Apart from postponing (sometimes for quite a while) the inevitable, do doctors heal, or do they create the right setting (incl. medicine and surgical changes) for their ’employees’/systems/patients to do their thing, the healing they want to ..?

[Edited to add:]
Oh by the way, yes I do realize there’s more to management of course, in particular middle management, on the side of ‘defending’ one’s department and getting sufficient and right resources, and stuff. Reporting, less so. As if that were the goal… Etc.
Upper management (sometimes dubbed ‘governance’, wrongly, as in this post) has more coordinating tasks, and more leadership to display, and a more mint picture to present to the outside – which includes showing the dirty bits ..! (think that one through)

InfoValue, at last ..?

Ah, finally, some depth shows up to have been developed in the field of infonomics. Very, very happy to see that; as you may recall from many previous posts of mine (and from before the very idea of posts, or blogging, or the Internet as such, did exist) I had an urge to find, or (help) develop, methodology in this direction. So there you have it, and there.

I’m happy for today. But will post much more on all sorts of bits and pieces that may spring to mind, in the near future. May even want to create a new category for this.
Anyone out there that already has more on the, say, more hardcore accounting side of this? Much needed!

And oh, of course:
DSCN4011
[Bridge, over as yet untroubled waters, Calatrava of course yes indeed, at Valencia]

(A category of her own)

By faking in all sorts of ways, she created a category of her own… the Yolanthie!
As in this, and this, and elsewhere

But hey, that’s how innovation goes, right ..? Trial and error.
[In this case, creating a completely new fantasy category…]
Also ruthless vetting, more than ever, though. Not really the benign advisory type – that would take some maturity, apparently not present in the immediate vincinity of starlets nor with the haters; gonna hate, out of lack of any own initiative, innovation, or self-worth altogether.

And here’s what you were waiting for; a real one!
DSCN7735
[Smile, Baltimore style]

Finally, real responsibility

..it’s getting real. A CEO has to take personal responsibility for an incidental security miss.

So, after all the talk about ‘top management being accountable, and you can’t just delegate that’, we see that the buck indeed stops where it should. Or, well, it stops there, too; probably, down the line some others may receive a slight dent in their bonuses (up high) or get fired on the spot (lower down). But, for once, the blame doesn’t run downstream all too cleanly until it reaches the bottom.

Oh how wonderful it would be if, if this would become the standard. Then finally those in InfoSec and ITSec would get the appropriate (!) budgets…

Oh my! I forgot to mention Cyber …! Just sprinkle it around in the above piece, to make it hip and modern as if I understand everything new in the world – or it proves I don’t, par excellence…
#ditchcyber / #wegmetcyber you know.
So, here’s your picture:
DSCN5083
[Hi, Ludwig!]

Account your blessings, or not.

Accountants of the certifying kind are still searching for their purpose. In life but moreover, in business reality.
Which isn’t too strange. They exist only for a couple of centuries, so at the speed of thinking they do, now is about the right time to do a second round of thought about their role.

First, a picture to freshen the mind:
DSCN1240
[Memento for outdatedness..?]

To start off, accountants [from here, I presume them to refer to the external auditors of the annual financial statements] were involved in checking all detailed financial data as well as seeing these represented in a standardised form into the annual statements. These annual statements had not very, very much flexibility, (so / because) any statement’s line item had a quite limited interpretation space. Sort of XBRL avant la lettre, in practice.
But since, data representation standards have complexified so much that there is no ‘standard’ anymore; interpretations mushroomed. Which is why exactitude of the summaries in the annual figures isn’t anything important per se anymore, and the fuzzy representation of performance is.

Along came many more types of accountants, of auditors even. Though some may narrowmindedly think all ‘auditors’ are actually the external certifying accountant type, they are dead wrong.
Continue reading “Account your blessings, or not.”

Maverisk / Étoiles du Nord