A Deeper Explanation of Neural Networks (Easy to Understand, Though)

Title: Comment on a Medium Article

Source: Surfin’

Link: https://medium.com/@nafrondel/you-requested-someone-with-a-degree-in-this-holds-up-hand-d4bf18e96ff

My Perspective: Back in the 1980s I took a couple of classes in neural networks.  It was, at the time, the logical continuation of my interest in AI: we had little machine power, but we had C and Pascal (not to mention FORT and Assembler) and that was going to be enough.  The first course was a three months jaunt into building a neural network that could recognize any letter in in the alphabet by either hearing the sound or looking at a printout/drawing of it.  This is 1986-87.  It was fun, but it also introduced me to the complexity of a neural network and how we actually have to teach computers.

I found the above article while going through rabbit holes on AI a few months ago, and I really liked it because it does a great job of simplifying that first foray into neural networks (which we use, in different forms, for machine learning and “deep learning” (marketing name for machine learning)).  It covers what for me became the next long search: to understand cognition and learning (still don’t get the most advanced concepts, but it’s a fascinating hobby and science).

I am hoping that by reading it and mixing it with my previous link on ML, you begin to both see the potential and complexity behind ML (and realize that for most vendors today, it’s just another marketing term).

disclaimers: this sh— stuff is hard AF, but the easy part is to understand how and what cognition works.  when i say most vendors are hyping and marketing, i mean no particular vendor — and of course, none of my clients  — but a generalized statement considering the complexity and, still, academic nature of true ML.  that first course i took? we barely made it past helping the machine learn both what a B was and how to recognize it.  today you can get that in an open source library, at worst.  c'est comment la vie avance…

The Next Decade in Digital Transformation

A knowledge summary is a semi-long to a long post that synthesizes positions, concepts, and lessons learned around a topic.  They consist of a mix of primary research with ideas and frameworks I built based on conversations and working sessions.

This knowledge summary will focus on concepts you have to know to succumb to embrace digital transformation in the next decade.

Understanding Digital

I wrote about digital transformation first in 2013.  Then an update later in the year, and finally a call to end the talk in 2015.  And I wish, you don’t know how much, we could do that.

But we can’t.

The problem is that digital has become the new crutch word much like e-anything or i-anything became the crutch letter in the late 1990s.  And we have seen decent value arise from that use: iPhone and eCommerce are the two most recognized iconic words (and concepts).

The purpose of Digital is to replace all that has to do with computer-driven or computer-assisted.  We have had computers since the 1980s in the organization, we cannot say “computer-driven transformation” as that ship has already sailed;  we talk about digital as if it was a magical, mythical approach to changing things.

But it’s not.

At the core, this machine-driven revolution is different from the one 30+ years ago in what is focused on – data.  That’s it, as simple as it can get.  This transformation is about data.

It would be more appropriate to say it was about information (a combination of data, content, and knowledge that enables companies to solve problems better in context), but — as I said before… i-anything was already taken.

Seriously, it’s not about data only, it could never be.

Data is the representation of an event.  What happened to whom or which, when, how, and where – it’s the pieces of information you need to recreate something: a customer signed up for a newsletter (what), they received the newsletter (what, when), they opened three of the five links (what, when, which), they did a search for further information (what, how, when), they bought a product (what, how, when, where), we shipped it (what, when, where, how), they received it (what, when).

That’s all data – every piece of that.

Data does not mean much without the other two parts of information: content and knowledge.

Content is the static information we (or someone else, if we believe in communities) created that describes an entity (product, solution, use cases, manuals, etc.).  For example, the customer got a newsletter with content – information about products, services, discounts, coupons, etc. and acted on that.  The data about the newsletter being distributed and the action remains, but without the content, we would not know why the client acted and we couldn’t track specific actions they took.

And that brings the final piece: knowledge.  Knowledge, if you follow my writings, is wisdom – the applied, contextual, intent-driven use of content.

And that is where the reality of what data can do comes together — if you combine data about events with content related to that event and intent and context in the form of knowledge – what do you have? the right information, at the right time, in the right place personalized for each person, optimized for each situation, and outcome-focused.

In other ways, you have the reason your organization is trying to adopt a business transformation strategy.

Becoming Digital

Now you know what digital is and why it’s the source for business transformation this time around: even though data has been around from the beginning, it’s the leveraging of data mixed with the right content and knowledge, yielding information, that makes it the basis for this new evolution.

How do you use it?

There are two slides that I use to show my clients and people I talk to where / how / why information is to use it.  First, where does information come from?  Big Data — rather Big Noise.

Check it out…

The concept of Big Data is a misnomer: how can something be “data” or be anything if we don’t even know what it is.  If one trillion IoT devices generate measurements, which ones are truly data and which ones are noise? If there are 20 millions tweets in one week mentioning your company name, product name, or someone in the company directly by name, how do you know which ones to reply to directly and which ones to safely ignore? What about corrupt or misspelled or damaged “data”? it’s not data until you structure it, until then is noise – simple.

Although the first reaction is always to store everything and “later we will figure it out” the volumes are becoming too large for that, and the need to use the data (and the speed at which data usability and usefulness decay) increases dramatically every day.  Storing for future use is no longer possible these days.  The next step is to filter the noise and create a signal.  If a plane that uses Rolls Royce engines generates over a terabyte of data per flight, and there are thousands of flights daily worldwide – which piece of “data” should be stored? used? discarded?  Filters make that decision.

Once the filters take their turn and the noise is selected as valuable, then we structure the information.  This is the decision as to what it is: data, content, or knowledge?  This is the moment the noise becomes structured and ready to be acted on.  The concept of digital transformation being about acting on unstructured data is another misnomer – how can something be acted on if we don’t know what it is? where it came from? what is used for?  once we filter the noise (create a signal) then we figure out what it is that we have, structure it, and store it in the proper place (or not, but that’s another post) and act on it.

These actions, the analytics or workflows applied to them, generate an insight – a something we did not know.  These insights are what we use to make decisions, to act, or simply to inform – via dashboards and reports.  These signals are also the components of information that will be used to power algorithms and AI components – and that will become the new filters for Big Noise.  The cycle of optimization based on what we know and what we learned is then complete and we have both insights we didn’t before, and data (and content and knowledge) stored.

Very different from simply capturing everything to figure out later what to do, no?

Being Digital

This section is the second framework I use with clients and in conversations when talking about what it takes to become a digital enterprise.  I must confess, this is not a new model – I first introduced it in 2013 – but have been working and optimizing it since.  Thanks to all of you that helped me along the way.

Once you get past the awesomely-designed majestic use of colors and boxes, you end up with six components you can focus on: four yellow boxes, and two blue boxes.  I know, mind-blowing.

The yellow boxes represent infrastructure, the purview of IT and Architects in your organization.  As your organization becomes more entrenched in the cloud (which by now is commoditized, so I am assuming you are either already there or on your way there) they will notice more and more a need to structure their approach.  The yellow boxes separate the different pieces they need to focus on:

1. cloud infrastructure, the core components used to run and interconnect everything
2. legacy access, because I am still proud of the code I wrote in the 1980s in COBOL and to be fair – a large number of you still use it and need it to run your organizations
3. interface connectivity, because in the age or mobile and IoT there is no longer a requirement but actually a necessity to be device-independent — this is not mobile-first, this is mobile-also
4. and to accommodate the hype of the day – AI (or advanced analytics, if you read my writings on the topic) with three specific outcomes: optimization, personalization, and automation.

I wrote about this framework before, the link is above and there is an update to it here as well, which is why I am just summarizing it here.  If you need more details read those posts, or contact me.

Once you configure your distributed computing architecture (AKA cloud) to operate in a model similar or comparable to the above, you will end up with two blue boxes – which is where the magic behind digital business transformation happens.

The information blue box is what we were discussing above in the previous section: how to find the right data, content, and knowledge to create the necessary information for every transaction.  The information will come from anywhere: devices, legacy data or applications, or AI and analytics engines.  The role of the information layer is to make sure that all information is considered, and the best selected, when crafting the response to an interaction.

The experience layer is the one where most organizations would love to have control – but they can’t.  The concept of building experiences for customers is archaic and, frankly, dumb.  This is not about understanding customers journeys or planning for them – or anything like that.  This is about understanding that stakeholders (notice it says customers as well employees, partners, and public) will interact with the system on their own terms, according to their expectations, via any channel, at any time, in any way they see fit.  Today they may have more time, tomorrow less – and they would appreciate a quick summary today, and more information tomorrow.  To accommodate these shifting expectations, each experience will be built ad-hoc by the stakeholder – provided they have access to all systems, information, and rules that apply.

This is the essence of being digital: building an infrastructure that allows any stakeholder to interact with any part of your organization, at any time, anyway they want, for anything they need.  if you can do that, you are “living la vida digital”

To get there, that’s what you need to do over the next decade; this is not a simple system purchase and deployment.  It requires extensive changes in all layers of the organization: people, process, and technology.  It also requires new thinking in governance and metrics.  and that — that takes a decade or so to complete.

Your move.

disclaimer: first things first, thanks to Jon Reed for the idea of using crossed-out text.  Since Sameer Patel does not like my parenthetical digressions, I am testing some minor ones using crossed-out text to see if it works.  if it does, all credit belongs to Jon who uses them to great effect (far greater than I could ever do) at Diginomica. I would reluctantly, but understandably, relinquish the use if he asks.  second, I also have to give credit to the kind folks at OpenText who helped me with the first chart a bit during a consulting session in December.  the other people, many, who contributed over time are also very kind and i am profoundly thankful – but they were more motivation and inspiration, OpenText tweaked the slide for me and let me use it.  third, no vendors are mentioned – but y’all know that i work mostly with them, so there’s a chance that some of this stuff shows up in something they do / use and i will gladly take all the credit for that.  ideas are mine, originally and follow-through, so feel free to yell at me for anything i got wrong.  if it works, not me – if it didn’t, me.  comment box is below.  thanks for reading.

Research Summary: Nine Starting Points For Digital Transformation In Manufacturing

R "Ray" Wang

Principal Analyst and Founder Constellation Research

R “Ray” Wang is the CEO of Silicon Valley-based Constellation Research Inc. He co-hosts DisrupTV, a weekly enterprise tech and leadership webcast that averages 50,000 views per episode and blogs at www.raywang.org. His ground-breaking best-selling book on digital transformation, Disrupting Digital Business, was published by Harvard Business Review Press in 2015. Ray's new book about Digital Giants and the future of business, titled, Everybody Wants to Rule The World was released in July 2021. Wang is well-quoted and frequently interviewed by media outlets such as the Wall Street Journal, Fox Business, CNBC, Yahoo Finance, Cheddar, and Bloomberg. Short Bio R “Ray” Wang (pronounced WAHNG) is the Founder, Chairman, and Principal Analyst of Silicon Valley-based Constellation…...

Read more

Marketing Dividends – Is it Time to Re-evaluate Digital?

Gavin Heaton

Founder Disruptors Co

Gavin is the Founder of Disruptors Co, one of Australia’s leading independent strategy and innovation companies. Gavin advises firms on strategy, marketing, product development and innovation. He is also a member of the Orbits program, the influencer network of  award winning analyst and advisory firm, Constellation Research. Gavin advises boards and leadership on innovation strategy, marketing, data-driven product development and corporate venturing. Gavin is on the board of Vibewire, a youth-led social enterprise based in Sydney, Australia and on the Technology Advisory Group for Good2Give – the workplace giving platform. ...

Read more

Accordingly, substantial investments have been made in a wide variety of technologies from CRM and data mining, to automation, analysis and beyond. In fact, Scott Brinker’s infographic on the landscape of marketing technology (2016) suggests that there were almost 4000 marketing technology solutions vying for your attention and purchase. With so many choices, it’s hardly surprising that marketers wonder where to start with the MarTech stack.

But Byron Sharp, Professor of Marketing Science at the University of South Australia says that the promise of digital marketing is unfulfilled. Or perhaps, we have over stated the role of digital at the expense of brand. This video segment by the Australian Association of National Advertisers (AANA) touches on these topics, raising interesting challenges for us all.

Marketing Dividends Episode 10 featuring Byron Sharp, ‎Professor of Marketing Science, ‎University of South Australia

The final episode ‎of the Marketing Dividends series presented by the AANA (Australian Association of National Advertisers with Byron Sharp) with Professor of Marketing Science, Director Ehrenberg-Bass Institute – ‎University of South Australia. Marketing Dividends is a content series on SKY NEWS BUSINESS aimed at elevating the profession of marketing and explaining the value marketing brings to businesses and the wider economy.

Now, there is plenty that I could argue with. There is a huge assumption that analogue marketing metrics are/were valid, and also that marketers are not following through on data, analytics and measurement of business value. But these are quibbles – because the most interesting aspect of this interview is the refocusing of marketing towards strategy.

In many ways, the pursuit of digital marketing and technology has seen us become reliant on tactics masquerading as strategy. We put some technology in place and think that the strategy will magically be enabled.

But this is never the case. As Byron reminds us, “We are in a battle for attention – for physical and mental availability … people [consumers] just don’t think of you enough”. Segmentation, data and technology alone won’t solve that problem – only a tightly threaded strategy and approach to execution will. And that means doubling down on your marketing skills. So don’t just re-evaluate digital – re-evaluate your team and yourself.

4 Memorable Digital Marketing & Sales Effectiveness Quotes

Director of Customer Programs, Constellation Research Constellation Research

Carole oversees the Constellation Executive Network, an executive community for innovative senior leaders who value strategic guidance from trusted advisors who understand the power of disruptive technologies. Carole brings over 15 years of marketing, strategic account management, and operations experience, specializing in high technology and professional services. She holds a BA from Santa Clara University and a MBA from Thunderbird, School of Global Management.     ...

Read more

Cindy Zhou

CMO Level Access

Cindy Zhou is a practicing CMO.  Prviously, she was Vice President and Principal Analyst at Constellation Research covering Digital Marketing Transformation and Sales Effectiveness. With over 18 years of practitioner experience in corporate marketing, product marketing, product management, and sales operations, Zhou has spearheaded marketing transformation at multiple technology companies. She advised Constellation’s clients on strategies to light-up demand generation, prove revenue contribution, and maximize sales productivity. Prior to joining Constellation, Zhou was Global Senior Vice President of Marketing and Sales Operations at BackOffice Associates, a global provider of data quality and information governance solutions to the Fortune 1000 and SAP SolEx partner. Before…...

Read more

Even if you are not a natural disruptive thinker or leader, thinking creatively about innovation and and the role that technology plays can be easily stimulated by seeing a quick tweet, quote, or hearing a joke. After a quick online search, I found inspiring innovation quotes from current leaders and spotted general technology quotes like these. 

Business Insider - 27 Insights About Creativity That Every Entrepreneur Should Read
Forbes - 20 Great Technology Quotes to Inspire, Amaze, and Amuse

At Constellation Research, our analysts share ideas related to their business and disruptive technology research. In that context, the meaning of each quote, especially for business leaders interested in these topics, can be more useful and relevant than a generic one whether it's a quote about technology optimization, the increasing changes in the executive office, which we call "the new c-suite", or digital marketing and sales effectivess among many others.

We update these weekly for anyone interested in non-mainstream disruptive thinking from our seasoned Constellation analysts and keep an archive of the entire collection for our Constellation Executive Network members.

DOWNLOAD THIS WEEK'S ANALYST QUOTES - DIGITAL MARKETING & SALES EFFECTIVENESS - Cindy Zhou |  VP & Principal Analyst | Constellation Research 

"#CEO & Boards want predictability in the business. Embrace your inner #datascientist." - @cindy_zhou #CMO 

"Command of #marketing financial metrics helps change the #CEO perception from cost to growth center." - @cindy_zhou #CMO 

"It doesn’t matter if #sales or #marketing generated the lead, what matters is closing the deal." - @cindy_zhou @constellationr #CMO #CX

"Every great #CMO started somewhere. Build expertise, become known for it, then broaden skillset." - @cindy_zhou @constellationr 

Connected Planning and Promising Account Based Solution Demonstrated at Anaplan Hub17

Cindy Zhou

CMO Level Access

Cindy Zhou is a practicing CMO.  Prviously, she was Vice President and Principal Analyst at Constellation Research covering Digital Marketing Transformation and Sales Effectiveness. With over 18 years of practitioner experience in corporate marketing, product marketing, product management, and sales operations, Zhou has spearheaded marketing transformation at multiple technology companies. She advised Constellation’s clients on strategies to light-up demand generation, prove revenue contribution, and maximize sales productivity. Prior to joining Constellation, Zhou was Global Senior Vice President of Marketing and Sales Operations at BackOffice Associates, a global provider of data quality and information governance solutions to the Fortune 1000 and SAP SolEx partner. Before…...

Read more