Job Killers or Partners?

More than forty years ago, I've made my first step into the "Blue Tower" of Austria's largest industrial enterprise, home of the "Electronic Data Processing" department. An algebraist by education, I was hired to establish a "CNC machining" software development group.

Automation or Augmentation?

After that first step, I've built software systems automating machine tools, robots, assembly systems…and finally whole factories. In the industrial evolution: static-->mechanized-->electrified-->computerized-->cognitized-->connected, I've started in the early stage of computerization.

Automated factories produced distinct items faster, cheaper, preciser...with much less people, but automators acquired new skills to leverage technologies. In the mid 80s I've managed about 100 mathematicians, computer scientists, manufacturing experts…

Our work has displaced workers, but also augmented the demand for skilled labor at the user as well the builder side.

Today, as AIs encroach on knowledge work, it's hard to see how humans will remain employed in large numbers.

Although this is a political issue, it's my strong belief that avoiding an antisocial polarity, future work needs to become more hybrid, life-long education needs to be emancipated from schools, teachers and curricula, weekly working time needs to be reduced and work life reframing the use of intelligent systems as augmentation.

What kind of positions can they take to empower, of knowledge workers? Demanding but not killing?

Super-Brain or Collaborative AIs?

25 years ago, I've conducted my first AI project and I know, the progress in AI is incredibly fast.
Since then I've managed machine learning projects for metal forming, ski production, paper making, printing, medicine…we've even made our own multi method machine learning framework…with crisp and fuzzy decision techniques, ANNs, kernel technologies…represented in a symbolic programming language.

Companies as Google, Amazon…want to create a channel between people and algorithms...we just need tot rebuild our systems by massive inherent parallelism, massive data and better algorithms they say…AI that is driven in large nets will cognitize things…our thinking will be extended with some extra intelligence?

Collaborating with smart machines, we can learn how to manage difficulties and even learn more about intelligence.

Agreed, but I talk about AI as a set of techniques, from mathematics, engineering, science…not just a ubiquitous post-human species.

I believe in the intelligent combination of modeling, adaptive calibration, simulation…with an intelligent identification of features. The storm of parallelism, bigger data and deeper ANNs alone will not be able to replicate complex real behavior.

We need to continue making knowledge computational and behavior quantifiable, but even more important, find new interaction paradigms between intelligent "Silicon" and us.

Innovations Supporting Career Steps

I've developed The Innovation Mesh as a diagnosis scheme assessing whether an innovation will work and sell. It adds emotional factors. I've introduced The Quant Innovation Spreadsheet and The Quant Innovation Mesh for the assessment of systems that require quantitative treatment and filled them in for UnRisk Capital Manager...with references to UnRisk Quant to explain the differences…especially through the lens of the actors. What makes the innovation thrilling is the dominant question.

But, with an augmentation mindset actors may come to see innovations as (sparring) partners in doing the difficult work and promoting their career.

In which possible steps?

Four UnRisk Ways to Augment Actor's Strength 

Let me construct an "UnRisk" developed from the UnRisk Technology Stack: the UnRisk Financial Language, implemented in the Valuation-, VaR- and Exposure/xVA Engines, the UnRisk Data Framework, Interface Builders and Deployment Services. it's built as solution and development system in one, covering a wide range of deal types, portfolio and scenario instances…all practical details mapped…for usage from the front office to accounting.

What possible career steps can an actor, a financial expert, consider…driven by the values they can most likely add?

Climb - a strategy of, say, a front office expert...heading higher management grounds, avoiding the plumbing, to become capable of more big-picture thinking striving for better risk-informed deal decisions. For her, UnRisk will play the role of a digital risk quant and manager.

Dive In  - a strategy of, say, a financial engineer structuring new deal types, funds…exploiting UnRisk's symbolic Financial Language and inherently parallel and platform-agnostic engines building new valuation systems and tools swiftly. For her, UnRisk acts as a library quant providing foundations that are comprehensive and bank proof.

Surf - a strategy of, say, a financial advisor…helping non experts to understand complex financial concepts. They may even head a position where they help their financial institution to find new offers for new clients and define competitive game rules…For them, UnRisk will act as a deal composer and simulator supporting smart user interfaces enabling simplicity but explanatory depth.

Leap Forward - a strategy of innovators, who want to build the next generation of quant finance products for their own markets. They will exploit UnRisk's Financial Language, Engines and Services, stretch and expand them in co-evolution with their product life cycles. They could even build UnRisk atop UnRisk.

A lot of brain work cannot be coded (machine emotions and machine thinking will be different)…actors need to decide in which direction they want to enhance smart systems.

If I once walk through the city with my digital aura attached…an information may pop up: "Herbert, hi, I'm your smart portfolio…I've engineered an instrument mix that fits perfectly to you risk appetite, the probability to win is…your log loss expectation is…my adaptive Kelly calculator recommends a revolving investment of x% of your wealth…with 97% confidence...".

It'll be still realized by preprogramming machines…when will machines learn to program themselves?

And we better don't think like machines to enable machines to think like us.

However, even if the path beyond automation goes to augmentation, there are big challenges waiting for the politics.

 Parts of this post's been inspired by this article in HBR, Jun-15 issue.

Quit What's Blocking Progress

I've just finished reading This Idea Must Die - Scientific Theories That Are Blocking Progress, the 2014 Edge question. What scientific idea is ready for retirement (so that science can advance)?

About 180 answers of visionary thinkers are collected in the book. The Big Bang was the first moment of time. Entropy. Infinity. Cause and Effect. Essentialism. Free will. Cognitive agency. Robot companions. Nature = select a few. NN Taleb takes down the standard deviation (again) and Emanuel Derman questions the power of statistics (I've read the books of both motivated by my engagement in quant finance).

Ideas that block progress

But what I'm interested in more general: few radical innovations are developed without first abandoning old ideas and systems. And some ideas do really block progress:

One size fits all - in-search-of-perfection seduces us to select frameworks, models, methods, programming languages...that fit "everything". But what we need is multi-model and -method approaches, hybrid, multi-paradigm programming…down to platform-agnostic systems.

Data Science replaces everything - the opposite is true…it's hard to work in the data salt mines and the danger of being fooled by noise is ubiquitous. Intelligently combining modeling and data driven methods is the right way to go.

Tightly coupled complex systems - A complex system can have untended consequences and tightly coupled means there is not enough time to react to them. Diversify and decentralize your system, make it agile by increasing its adaptability and optionality. Make intelligent independent system components that are accurate and robust, but flexible. Let them co-exist and co-evolute.

Product = technology - programming languages, engines that implement them, data frameworks, deployment systems, clouds…are technologies enabling many transformative developments of products,

AI = neural nets - first, the expert system thinking of AI died, then the idea of "Artificial Life"…now we think: because our brains have neurons AIs are well represented by neurons too?  IMO, AIs are built of a set of techniques of mathematics, engineering, science…not a post human "species". Artificial neural nets will be of great help in this multi-factor approaches. Behavior must ve quantified and knowledge must be computational.

Symbolic (exclusive) or numerical computation - both…the mathematical discipline is asymptotic mathematics: us symbolic constructors and builders to get models of good nature for numerical solvers and use symbolic computation for dynamic result interpretation.

In short, forget the (exclusive)OR, believe in the AND effect.

Risk's Not Easy To Understand

I like reading the Eight to Late Blog. It's at my "As Read At" list, I walk through almost daily.

The latest post is about The Risk - a dialogue mapping vignette - about a workshop on the theme of collaborative problem solving using dialogue mapping as tool with a special notion: Issue Based Information System.

Risk or danger?

It's insightful to see how the tool helped creating the map by team discussions evolutionary...but there's one thing: "Schedule is too tight" is not a risk, but a danger. How I turn it around, I do not find a realistic opportunity in it.

Optimal risk revisited

"Build a technology stack to master tight schedules of future projects..." is a risky project. If you run out of financial capabilities before anything applicable is done it kills…But if it's done it creates a lot of opportunities: do things you couldn't do without, applying it in fields you haven't thought about before, more flexible licensing, pricing and return valuing…and options: minimum viable releases for minimum viable usages for minimum viable clients…in an evolutionary development approach.

Risk is a subject of common misconceptions. Remember, managing risk means arranging things that make opportunities and danger a positive contribution. If things are quantifiable, you may be able to optimize it.

Cutting off opportunities from "risk" creates fear

The workshop was a great explorative learning experience…I've no doubt. What I want to point out: the deep understanding of the principle factors influence the mapping of the conversation about them.

The philosophical background of this is that things "change" in discussions if their weave of formal representation and content change. If risk is bad becomes the common sense (a formal "law"), we stop innovating.

It's done before? It's too early? If we do it, it may be the last time? What will the market say?…are all fruit of risk aversion fed by mistaking risk and danger (loss).

My Personal Navigation System

This post has again been inspired by Shawn Coyne's Blog

The process of assessing an innovation can be overdone. Especially when it's driven by essentialism or perfection. There's no metrics and things change rapidly…and many innovations are about the new for the now.

So, if I walk an innovation through a scheme, say, I'm Innovationmeshing systems like Opexar, after UnRisk,  I do most probably come to a point where I hit a barrier. It's a bad "I shouldn't have started this" moment. Stuck in an underwood? Or just weeds?…What to do? Shall I just stop go back and seek a new path? Or fly out?...Shall I improve The Innovation Mesh and start again?…or just "burn" the barrier down?

Real option valuation isn't identical with financial option valuation…real options and their underlyings are usually not traded…

It's so many possible directions, where am I on my assessment map?

My Personal Navigation System

Why I'm innovationmeshing UnRiskOpexar...?  

They're quant systems that are solutions and development systems in one…First I want to test my tools and then I want to use the insight and results to help innovators in the field of business and finance leverage their businesses.

What shall reference Innovation Meshes be fore?

They're results of the assessment tool applications and shall show innovators how those, who attack complex problems in certain fields make thrillers…building systems that are solutions and development systems in one.

With reference I do not mean "popular" but "great"

 Who do Innovation Meshs benefit?

Innovators, innovation marketers, marketing advisors (my competitors), investors, public innovation program developers, politicians…Understanding innovation is a skill.

Where to go?

First, I'll do more lab work assessing existing innovations and publish results in agreement with the innovators. Then, I'll improve The Innovation Mesh schemes and tools and publish further insight and results.

When will it be finished?

I feel that I'm quite close focusing on the segment of above: business and finance / quant innovations that are solutions and development systems in one.

But, because I'm confident that the principles are valid for other quant innovations (in engineering, automation…) there's more work to do. So much, that I can't do it alone.

Innovators - Racers At Critical Paths Or Master Builders?

Innovation projects are complex - people are working on many activities that are dependent on others.

Innovation projects have a Critical Path?

The project network - you have the work break down into tasks...a duration for each activity, dependencies and milestones? The longest sequence of tasks in the project network that are dependent and cannot be compressed is the Critical Path.

But do the organizations always know what that means?

The problem is, innovators are usually excellent and shall be asked…but people on the critical path must not be interrupted…because delays on the critical path are delays at the very end of the project?

A little analysis tells us who is on the critical path. And they should wear "I am on the critical path" T-shirts or caps…and team members without should ask them: "can I do anything for you?"


Innovation projects must not be linear

As we know from The Innovation Mesh they span "objects" in a multi-dimensional parameter space.

In particular, innovations that require computation have various objects of desire, a clock, flows, events, transformations, massive information requirements…most of the problems suggest a functional decomposition and the bottom up composition of functions and tasks.

Innovators think like master builders?

Walking an innovation through The Innovation Mesh needs lab work on the micro level. If this is done, the macro view emerges…but you may need to go back...

Doing this it comes to mind that key developers of quant innovations should think like master builders (of, say, a hotel complex or even parts of a city)

Apply a general procedure:

Questions - What are the desires, requirements, needs? What are the conventions, rules, state of the art? What are the controlling ideas, models, methods and critical implementations? What are the constructors, building blocks?...finding questions guides us towards ideas.

Frog and  Bird Views - iteratively zoom in and analyze the concrete details and zoom out and see results in the context of behavior and profiles.

Elimination - don't forget to eliminate features and capabilities that will knowingly lead you to unclear outcome. If you have eliminated too much you will find a better way...

Make - the concrete implementation is a a way to do practical analysis and verification. If you have great building blocks the final solution will be result of an evolutionary approach.

A critical path gets another meaning when applying evolutionary prototyping...

UnRiskers unrisk

Programmers program, teachers teach, swimmers swim…

Do UnRiskers unrisk?

Yes, but not in the sense of removing risk blindly. It's about how to make danger and opportunities a positive contribution…always in the regulatory framework.

UnRiskers know that there's optimal risk and they know that diversification does not always work and they also know that other "deterministic" optimizers, like the Kelly Criterion, need some deeper analysis.

And they know that hedging becomes even more difficult if more info of the individual market environment is known. To unrisk might require reducing sensitivity or volumes - but that are often conflicting objectives.

What investment and risk mangers enjoy using UnRisk: getting insight by comprehensive scenario analysis...the ability to shift multiple market and risk factors simultaneously…some predefined and automated, like in the UnRisk VaR Universe.

The robust UnRisk Engines deliver the values, risk spectra, cash flows...blazingly fast, delivering results of the required massive valuations in time. Portfolio across scenario simulations are performed automatically, inherently parallel...

The complex new regulatory regimes of central counter party and central clearing require the massive calculation of value adjustments, capital storage requirements (CVA/DVA/FVA…KVA), initial margin…UnRisk Exposure Engines need much more portfolios scenarios…

You may end up in multimillion single valuations to get support for a few decisions.

A bigger picture of risk

To understand the possible butterfly effects of, say, regulatory regimes you need quantitative treatment to a maximum extent…many results that can be aggregated, evaluated, visualized...

It's what UnRisk has in mimd: to unrisk...get a bigger picture of risk - not just remove it.

When the regime changes you need to unlearn, because the rules of risk management have changed. And your quant finance system need to manage the change.

Understanding risk you need to understand money, duality, boundaries and optimization.

The controlling idea behind "unrisk"

Our abstract concept is quite simple: nested instrument groups and scenario groups moving over time with events…That empowers structuring, portfolio, scenario and test building.

To unrisk means understanding the sensitivities of all relevant factors and optimizing risk in regimes where the "conditions" are known and relatively safe and remaining dangers are cleared out.

To unrisk means also to analyze the tail risk…calculate expected shortfalls, apply power laws to get distributions...

There's no such word in the dictionaries…but it got a kind of operational semantics by using UnRisk. 

UnRiskers unrisk, as UnMarketers unmarket… 

It was risky, when we decided for trademarking UnRisk, 2001. 5 mathematicians sat together deciding. The owner, the CEO, the product manager, the key developer and myself (business developer). It wasn't me, who proposed it...but I knew immediately: it  will contribute to a changed view into risk. A mathematical view.

It took me 14 years to understand that unrisk also means the thrilling story of managing risk by UnRisk.

Optimal Market Risk

In my previous post I've introduced RCD Marketing. I've written about the two-sidedness of risk Here.

Let me explain, why this will influence my future decisions in innovation marketing (advice).

The risk of a blast furnace

Iron is produced from iron ore, coke and additional materials (magnesite, decrease the melting pot, control viscosity ...). Among the processes of iron production, the blast furnace process is still the most prominent one.

A modern blast furnace produces up to 5 mio tons of iron / year and can be operated continuously for up to 10 yeas before shutdown.

To control the process you need to model various phenomena, like flows of iron ore and coke, liquid due to melting, gas - from bottom to top, energy - heat conduction and convection, mass and chemical reactions.

As the iron ore sinks it is indirectly reduced: Fe2O3 --> Fe3O4 --> FeO --> Fe (a process of improving material quality for the purpose)

BTW, the typical sitze of the problem: 40.000 spatial unknowns per scalar component of  20-30 unknown functions (temperature, max concentrations, velocities). The equations are of the reaction-convection-diffusion type and providing the required accuracy at speed you need to choose the mathematical solvers carefully.

To maximize the output you want to operate the furnace ambitiously but you must avoid that the liquid iron and slag erode the lining too early or, worst, melt through the shell of the furnace (if this happens, you better run).

With this respect, the risk of producing iron is a cost for securing the process, but also a strategic asset…it can be quantified, because of technical and economic modeling and simulation….consequently, optimized in the mathematical sense.

The risk of life

A good life? There is a widely expressed emphasis on the maximization of a benefit (health) and not the optimization of a risk. How to enjoy life, but avoid severe diseases? How much sport, how much mobility, how much interaction, how much celebrations, how much wine and dine…. how much medication…?

Yes, it is much easier to do risk management in quantitative fields - if dangers and opportunities are not quantifiable, we have limited ability to control them.  But you're lucky if you're able to make danger and opportunities a positive contribution.

The risk of marketing

It's quite has the dilemma in it: is marketing risk a cost for securing a business or finding the optimal input fraction of an overall financial potential. This are the poles of the dilemma. Especially related to innovation marking where uncertainty dominates: predictability vs adaptability, control vs agility…?

I've introduced the reaction-convection-diffusion metaphor to marketing because it helps to optimize the market risk qualitatively…you put (the story of) your innovation into the pot (market)…it will

  • react with worldview, states, solutions…in market segments related to its change potential and requirements
  • circulate by convection at borders of segments with gradients in expectations, knowledge, persuasion
  • diffuse through the market promoted by multiplying actors

Marketers can observe the "heat and speed" and control the process by adjusting access, information, education…avoid that the "pot" freezes the story or boils over with negative reactions.

Your innovation works and sells in principle…The Innovation Mesh tools helped to asses it…the reaction-convection-diffusion marketing metaphor will help optimizing market risk...

Reaction-Convection-Diffusion Marketing

The reaction-convection-diffusion model is one of the most frequently used models in science, engineering…and quant finance. It describes how the concentration of something distributed in a medium changes under the influence of reaction, convection and diffusion. It's a partial differential equation.

In a nutshell, reaction is a process that results in a conversion of something, convection refers to a movement of something in a medium, diffusion is the movement of something from a situation of high concentration to a situation of a uniform distribution. Obvoiusly, the names come from phenomena in physics, chemistry…but, some derivative pricing and risk models, like interest rate models, are represented as reaction-convection-diffusion equations as well.

The big picture of marketing

I borrow this principle…for marketing an innovation project to (innovator's) organizations, partners…clients.

You have an idea of an innovative project…you need to convince peers, investors, potential clients…how can you get them to understand, accept and integrate? Its marketing. And it's abut answering key questions.

Reaction - how much does your innovation require behavioral change? To what extent are those changes in disagreement with the worldview of your target groups? Is the story of your innovation (not so much the facts) so compelling that you can minimize the mismatch that effected the acceptance of the change?

Convection - is your innovation demanding to be talked about? Do your ideas spread with your target groups without your interventions?

Diffusion - do you bring your core ideas to early adopters and empower them to to move through the organizations to the majority?

No, I don't strive for a marketing theory of everything. And yes, in the details internal and external marketing need different actions…and no, I'll not try to create a RCD PDE model for innovation marketing.

But, I want to persuade you to walk through this questions carefully, before you decide for measures of operational marketing (down to promotion).

This post has been inpired by Seth Godin's marketing to the organization