Wednesday, July 1, 2020

Context: The Connecting Clues for Data

The “database of now” demands a quick understanding of data, particularly in context. There are many opportunities in understanding or misunderstanding data in terms of the contexts they participate in, or are connected to, or connected at the edge of a data neighborhood. Because each context has its own unique vocabulary, you can see the opportunity for misconnects in meaning by not understanding the full context of any statement or set of proven facts.

If someone says, "I like the blue one", how can you evaluate what that means? If it is a swimsuit on the beach, it means one thing, if it's a lobster from the same beach, that means a totally different thing. Context is what gives data real meaning. There are three primary forms of context that help understand the true meaning of the base data. One is the real world contextual meaning, the other is the contextual business meaning, and the other is the technical contextual meaning. Obviously, finding meaning in big or monster data is a challenge, but that difficulty increases as the speed increases, particularly if the data is hard to manage or access.


Figure 1 Representation of Interconnected Contexts.

 Real-World Context

Data has meaning in terms of its definitional domain. When you mention "blue", usually comes from the color domain. However, in the world of mental health, it means a kind of feeling or mood. So understanding the base context in which a data element exists is essential. If blue is associated with a human context, it could be physical and mean a lack of oxygen. It could also mean that the person is adorned in something blue. This is usually cleared up by understanding the base subject or entity that the data element is associated with by having a precise name, meaning, and basic subject area association. Underlying meaning can be tricky when just looking at the data value alone. Having proper meta-data and associations is the ideal solution to this problem.  

Business Context

The contextual areas that relate to business fall into three basic categories of meaning. Every data item or group of data items needs to be viewed in terms of the context they are being viewed in or from. Internal contexts where the vocabulary is understood in the context of the internal organization and defined within a particular organization. These internal contexts usually revolve around the organization and skill definitions. There is also the external context that represents the outside world irrespective of the organization itself. The third context is where the outside world touches the internal world. Listed below are the typical contexts in each of these categories:

Common External Contexts:

Communities, Brand/Reputation, Public, Legal Frameworks/Courts, Geographical Regions, Countries, Local Culture, Governmental Agencies, Industries, Dynamic Industry 4.0, Value Chains, Supply Chains, Service Vendors, Markets, Competitors, Prospects, and Competitors Customers.

Common Internal Contexts

Organizational Culture, Goals, Constraints, Boundaries, Actual Customers, Products, Services, Suppliers, Employees, Contractors, Departments, Divisions, General Accounts, Contracts, Physical Infrastructure, Technical Infrastructure, Properties, Investments, Intellectual capital, Business Competencies. Knowledge, Skills, Patents, Success Measures and Statements

Common Interactive Contexts:

Marketing Channels, Advertisements, Customer Journeys, Customer Experience, Loyalty, Satisfaction Scores, Processes, Applications, User interfaces, Websites, Webpages, and System Interfaces.

Technical Context

Data must also be understood in terms of physical contexts, limitations, and potential lag times. Data sources need to be understood in terms of their currency and ability to be integrated easily with other sources. While many views, interactions, and integrations work well at the logical level, physically, they may not be ready in terms of near-real-time capabilities, transformation potential, or performance levels on or off-prem. While meta-data may exist to understand possible joins and combinations, executing them fast enough to be useful in multiple business contexts may not be possible. The physical data types and file storage mechanisms may not be conducive to the demands of new usage scenarios. New low lag databases that are near real-time will become the standard, going forward.

Net; Net:

Data, information, knowledge are quite dependent on the context(s) they participate in or the perspective they are viewed from. Often Knowledge worlds interact; therefore, meanings can overlap and connect in ways that are essential for ultimate understanding, manipulation, or utilization. Knowing the context of your data is absolutely critical for leveraging understanding.  All of this is happening at greater speeds approaching the “database of now” speed necessary to make critical decisions, actions, adjustments or improvements. 



Tuesday, June 30, 2020

Generative AI+ Art is Gaining Momentum

I thought a post on generative art might be in the interest of all things AI. This kind of art is leveraging AI, algorithms, randomness, programs, and humans to create exciting and beautiful art. As you may know, I now collaborate with Fractal Software to develop compelling and award-winning artwork. In fact, some of my fractals are my best sellers. I have a great friend and fellow artist, Bob Weerts, who is pushing this collaboration even further. Below are two of his early generative pieces:












Bob employs lines as his fundamental stylistic element and incorporates a chance in determining line length, density, and color. He cedes some control over the work's final outcome to a process enabled by Software he's written allow the piece to "emerge" over time. He plans to let the Software take more control of these emergent pieces over time, letting AI/Algorithms expand some range. I find his early pieces quite pleasing and interesting already.

One source of Bob’s original inspiration is Casey Reas "Process Compendium," which, among other ideas, explored a synthesis of the Complexity Science notion of “emergence” and Generative Art in the early 2000s. An example of Reas Compendium work is below: (Click Here for Other Examples).

Reas is an internationally admired artist, but perhaps best known as the author, along with Ben Fry, of the graphical sketching too called "Processing," which is widely used in the domains of Art, Design, and Media. 

The significance of the generative art trend is perhaps exemplified by Christie's record of $432,500 sales of "Portrait of Belamy". The image is one of a series created by a group of young French students collaborating collectively as "Obvious".  Obvious borrowed heavily from open-source Generative Adversarial Network (GAN) algorithms specially developed by a then-high school graduate Robbie Barrat but originally conceived by the AI researcher Ian Goodfellow. This has the ball rolling, and there is new momentum under the "GAN" movement. Generative adversarial networks (GANs) are algorithmic architectures that use two neural networks, pitting one against the other (thus the “adversarial”) in order to generate new, synthetic instances of data that can pass for real data. They are used widely in image generation, video generation, and voice generation.

GAN's potential for both good and evil is huge because they can learn to mimic any distribution of data. GANs can be taught to create worlds eerily similar to our own in any domain: Images, music, speech, prose. They are robot artists in a sense, and their output is impressive. But they can also be used to generate fake media content of often called "deep fakes." 

Net; Net:

AI Generative Art is quite striking. Since the whole field is getting more towards AI and less from the artist/programmer, we can expect some exciting results in the future. I will likely pursue a more intimate collaboration with all kinds of generative art going forward. Keep your eye on Bob Weerts as he is a creative guy seeking this edge faster than many other artists.  

 

If you want to see my works, check out the fractals section here 

If you want to know more about my collaborations with Software to create, check out this post 

Read about more right-brained AI by clicking here 


 

 




 


Tuesday, June 16, 2020

Exploring Data Delivers

We hear about organizations mining data looking for benefits nearly every day now. Just like the prospectors of old, people are trying to mine gems out of the big patch of ground under their claim while searching adjacent areas. The case studies are abounding, so the appeal is strong. These mining efforts are this really paying off so how should one go about it?  Just start digging a big hole and hope for the best? With all the buzz around data mining and process mining, there are some proven paths to successful mining. 




Identify the Benefits of Data Mining 

It's pretty easy to justify the mining efforts on the promise of benefits today because there are so many success stories floating out there. The typical benefits that keep repeating include improved decision making, improved risk mitigation, improved planning, competitive advantage, cost reduction, customer acquisition, customer loyalty, new revenue streams, and new product/service development. The crucial step here is to find the benefits that will resound in your organization and situation. These days organizations are dealing with a multitude of challenges from plagues to politics. It is always a win to save costs, but there has to be more to it to create a compound set of appropriate benefits needed to justify mining efforts. 


Scoping Efforts Properly Delivers Better Results

While we all believe that data mining has the potential to improve and even transform organizations, the amount of data to mine is growing by the second and the number of advancements in making data smart is expanding. It's not difficult to understand that the majority of organizations are struggling to find the right strategy or solution. The first step is to discover where there is significant potential like miners do by drilling boreholes to discover the potential in the ground. That means organizations will have to sample areas of data that promise potential. To that end, many organizations start with process minging because it promises cost and time savings that often improve customer experiences leveraging smaller scopes. For those organizations that wanted an outside-in perspective, starting with customer journeys and large scoped processes that cross system boundaries have been quite successful. 


Incremental Learning is Essential to Continued Success

Starting small and expanding as success allows seems to be the most common model for mining. What is really popular at the moment is to use mining to find opportunities for more automation. Savvy organizations will look at adjacent systems, organizational units, and contexts. Feedback loops and iteration will teach the best lessons for mining results. Alternative visualization techniques such as timelines, animation, and limits do also help the learning process. Some organizations will also combine the visualization with discrete simulation to test alternative outcomes.   


Net; Net: 

If you are not practicing focused data mining and looking for productive patterns in your ever-growing data inventory, you are missing many opportunities. As successes emerge, the more savvy organizations are looking to widen their scopes and using approaches that cut through the jungle of their organization. Mining is here to stay and brings a valuable set of methods, techniques, and tools to leverage for organizations looking to thrive under all conditions. 








Tuesday, June 9, 2020

Organizations are a Jungle of Journeys

The simple idea of selling a product or a service for a price to make a profit is still the underpinning of most organizations. Still, it's gotten more sophisticated and intertwined than even five years ago. Many organizations participate in broader contexts like value and supply chains while dealing with dynamic change and emerging scenarios are driven by geopolitical or environmental trends/events. For continued organizational health, organizations will need to understand the journeys that exist and interact in their footprint of impact, learn the levers that can adapt their jungle to changing and in some cases, practice the response to the emergent conditions of "NOW." Listed below are the typical journeys that organizations need to participate in or manage in no particular order of importance:



Customer Journeys

The journey that a customer takes is a crucial journey to manage as it defines your organization's contribution to that journey that leaves an indelible memory of good or bad for all steps involved. It is essential to understand the customer's real journey, not just where a customer might touch your organization. Getting customers to be attracted and stay loyal to your organization depends significantly on your understanding of their real journey, not just the optimization or automation inside of your organization for cost savings. 

Work Journeys

Work arrives, gets assigned and moves through your organization, and is the key for cost and timing outcomes. Understanding where work gets stuck, deep in the innards of your organization, is essential for cost optimization and customer satisfaction improvements. It could be a competency skill deficiency, a data deficiency, an overburdened shared resource, or just a situation never contemplated for in the work design. These are some of the thickest vines in the jungle.

Employee Journeys

Employees are some of the most critical and expensive resources an organization manages. Making sure their time is optimized and used correctly is crucial for resource leverage with optimization reasons. Concurrent with employee participation with various journeys, they must be augmented and have enhanced/expanded skills. Assistance may occur through bot augmentation or knowledge turbocharging, but investment in employees is the often forgotten sub-journey. Lack of investment in employees is an easy way to lose in the long term.

Product/Service Journeys

Every product or service must be designed with the greatest of care and the best knowledge/ skills available. The journey from design to production should be planned, managed, built, and tested with the greatest of attention as they are often the competitive differentiator along with the customer journey and experience. Organizations tend to be very good at these kinds of journeys except when they become out of touch with trends or their customers, partners, and employees. 

Infrastructure Journeys

Organizations have to build and establish the infrastructure necessary to support the business. Service software has to run somewhere and needs to be built/maintained and supported by infrastructural software. These are part of the infrastructure that must be carefully and made promptly and retired if necessary as time progresses. The support will need to be built, maintained, or outsourced to other organizations if it is a product. Managing the portfolio of infrastructure during the building and maintenance periods are journeys to manage.  

Capital/Funds Journeys

Organizations are usually very concerned with money, how it is raised, how it's used, and what becomes of excesses or losses. While these journeys are better established and repeatable, they often try to dictate the level of investment in the other journeys. Visionary management will satisfy short term results expectations along with building for the future, thus fund incrementally in various journeys. Having a proper governance journey or two is essential for the investors. 

Community Journeys

All organizations participating in physical and logical communities that can affect them positively or negatively, the reputation and the operation of an organization. As organizations join in legal frameworks, the best are necessary to plan and execute the journeys that fit those contexts. The results will affect the kind of outside direct or indirect governance for organizations and may set the policies or rules for other remarkable journeys or processes.  

Net: Net:

Each journey must be thought through and managed collectively and individually. Traditionally only portions of individual journeys participated in digital optimization or automation. For organizations to thrive, these journeys need to be served digitally from one end to another. The interaction between these journeys will show where organizational friction will occur over time. Also, the interaction within these journeys must be orchestrated in the context of continuous foresight with emerging expected and unexpected scenarios.

The good news is that new digital business platforms(DBPs) are emerging to integrate digital functions to service journeys better. Some will help with process fabrics; others will manage the intelligence well for better decisions, or reaction/guidance; some will manage data integration, and still, others will work at the edge to manage emergence.

 

 


Thursday, June 4, 2020

Delivering Success with Smart Data Streams

It is becoming clear that AI will be a critical competitive differentiator for organizations, industries, and even countries. It is also clear that many are looking for success stories to leverage into learning opportunities. As AI embeds its intelligence throughout organizations, the sophistication of the data usage will increase to a point where traditional data approaches will need to extend to include real-time data streams of images, videos, speech, events, and operational data. It means that new data approaches will be necessary. As AI gets more sophisticated at speed, its hunger for complex data becomes insatiable. As organizations learn to leverage AI, emergent problems can now be attempted. You can find strong case studies of emergent AI acting on data streams by clicking here.



Leveraging AI Starting with Machine Learning (ML)

Machine learning allows applications to learn from the data in order to make better decisions at speed. There is significant value in creating predictive applications that can smartly select smart actions that meet or intercept emergent data from multiple and intersecting contexts. This iterative learning and improvement cycles are driven by emergent data, shifting goals, and guardrails that are invaluable for organizations that want to stay in step or ahead of their market place and constituents.

Intermediate AI Applications Leverage Smart Streaming

As AI gets more sophisticated in its learning ability by applying deep learning and even cognitive thinking leveraging interpretation, recognition, scoring, intuition, reasoning, and judgment, the hunger for faster multiple data sources will grow. Streams of complex and evolving data will need to be utilized in solving both static and emerging problems.

Putting a Premium on Emergent, Fast and Agile Data Sources

Looking to the past is valuable, but today's demands require organizations to get in front of business events, constituents, and competitors. The data sources will include traditional and non-traditional data such as voice, video, and images. The speed and mixes of data types and sources will be dynamic and agile. Instant integration and transformation will be the norm to satisfy prediction and intelligence needs fueled by AI and analytics.

Net; Net:

AI is gaining momentum and is taking on predictive applications that leverage fast and agile data sources. As AI migrates to the edge over time, the notion of fast streams of event and pattern data will grow along with traditional big and fast operational data. Organizations that want to thrive and capitalize on leveraging AI and smart streams will get ahead of the curve by learning from successful implementations. Please click here to access an E-book for some impressive case studies that leverage AI-enabled smart data streams.

 

Click here for the E-Book entitled "The Future Starts Now" subtitled "Achieving Successful Operations of ML & AI-Driven Applications."

 

This blog and this breakthrough E-book are sponsored by MemSQL(an agile real-time database).

 


Tuesday, June 2, 2020

AI Devours Data!

Those who have worked on Machine Learning (ML) projects know that ML requires a large amount of data to train the resulting algorithms. Some would say you can never have too much data. There is usually a correlation between the amount of data and the sophistication of the resulting ML model. This data hunger is only going to get more intense as AI progresses towards new benefit pools while leveraging more sophisticated AI capabilities. Since there are other contributing trends bedsides the sophistication of AI, the question looms for organizations is, "do they have the right data to fuel successful AI efforts?" If they don't have enough, should they inventory more in anticipation of the AI feast?




Figure 1:  The AI / Data Continuum

It’s not likely that all that big data that organizations have been hoarding is the correct data, but understanding where AI is going will give an organization a "leg up" on culling and collecting more of the correct data as AI progresses during the next decades.

The Progression of AI Changes the Data Game

While ML requires significant amounts of data to self-modify its behavior, the appetite of AI increases quickly as the sophistication of the AI capabilities increase. There is a big step from machine learning to Deep Learning (DL) in that DL requires much more data than ML. The reason being that DL is usually only able to identify concept differences with the layers of neural networks. DL determines the edges of concepts when exposed to millions of data points. DL allows machines to represent concepts via neural networks as the human brain does, thus allowing more complex problem-solving. AI can also work on fuzzier problems where the answers are more uncertain or ambiguous. These are typically judgment or recognition problems that can extend to the creation or other right-brained activities. This again requires more data, which in some cases may be emergent or real-time in nature.

The Shift from Data-Driven to Outcome Driven

As AI moves up in the sophistication of the problems its assists or solves, it will become data-driven and goal/outcome-driven. It means that the AI may request data on the fly that it needs to solve a particular problem or make a specific deduction, thus complicating data management. It may involve the interaction of inductive data-driven portions of a solution with the deductive needs for data based on a hypothesis to reach a target. This kind of dynamic interaction is needed for outcome-oriented problems. It is much different than just interrogating the data looking for interesting events and patterns. Decision driven approaches fit right in the middle of these two distinct approaches. Some decisions are operationally focused and improved through matching data with outcomes. More strategic decisions will pick up on both inductive and deductive approaches. This is just another demand channel to boost data usage.

The Shifting Problem Scopes Impact Data Needs

The scope of AI solutions are will typically start narrow and move to wider scope over time, thus requiring more data. Complex solutions typically target more than one answer and will require more data to support the tributary solution sets, contributing to a complex/hybrid result. As the scope of decisions, actions, and outcomes span more contexts inside and outside an organization, more data will need to be obtained to understand each context and their interactions. Each of these contexts could be changing and morphing at different rates, therefore, requiring more data yet.

Net; Net:

It's clear that more data will be the hallmark of AI-assisted solutions. The data appetite might come from more challenging problems, the better leverage of advanced AI/analytics, or growing end to end value chains. One thing is for sure. Organizations had better get ready for the new world of “AI/Data Interaction”. It could change or extend data management policies, methods, techniques or technologies.

 

 


Thursday, May 14, 2020

Dealing with Emergent Data


The recent and ongoing battle with COVID-19 has raised a goodly number of issues in and around getting surprised, arguments around emergent data, and slow/appropriate responses. The lessons learned so far are pretty rich, but I think there is more to discover. Scenario planning seemed to be lacking, the early warning systems seemed to have broken down, responses seemed slow and unpracticed once the denial hurdle was overcome. This was not a "Black Swan" event, so why did this pandemic seem to throw a monkey wrench into humankind's systems and processes? It may a bit early as all the dust hasn’t settled yet, but there are some obvious conclusions even now. What can we say now?





Scenario Planning


While pandemics are a practiced and expected scenario, the level of detail in this kind of scenario was tested in new ways. There were new discoveries in how trends in supply chains were working against us. With longer supply chains for medical supplies and equipment, the stress put on existing supply chains because of panic buying of many items, including food, and how do deal with rescinding demand in finely tuned supply chains. The detailed scenario planning and modeling really seemed off the mark this time on a worldwide basis. There were also some emergent geopolitical effects not completely thought through for sure. Scenario planning needs to handle models with more emergence in a fine-tuned fashion. Businesses and individuals need to up their game in this arena as well.

Early Warning

It is not surprising that less than effective scenario planning would lead to missing emergent data that was not expected, but as the emergent data morphed/changed, there were shadow events, signals, and patterns that took longer to recognize. Early warning needs to be able to recognize event patterns that go beyond expected events. These events and patterns need to participate in more complexity theory and real-time recognition that understands emergence in complex and interconnected systems and supply/value chains. This starts out at endpoint detection and merges with associated and real-time unassociated events to create emergent patterns. Agents/bots sniffing at the edge and event responding in the case of known situations and contexts is minimal, but merging and emergent complexities need to be tested, and models/scenarios need to be updated in near real-time.

Appropriate Responses

It seemed pretty clear that many responses were unpracticed, emergency stores were over-whelmed and tactical responses were being invented on the fly. Since appropriate responses are dependent on in scenarios and early warning, the compounding effect on responses were evident. There were some pretty impressive examples of human creativity/inventiveness and sacrifice to make up for the deficiencies, but can we plan on this always happening for the good of all? We saw governments taking over supply chains, people quarantined late and long, and decisions walking the razor edge between mortality and economic suffocation.
  
Net; Net:

We can do better.  We have to do better as more negative scenarios are emerging as nature deals out an accelerating frequency of earthquakes, hurricanes, volcanoes, pandemics, regional famines, and shifting geopolitical events. I suggest we invest in Emergent Data Recognition (EDR) tied to improvements in scenario planning and practiced responses from a strategic and tactical perspective. There are lessons in emergent data that we have to be ready to leverage. The more prepared we are, the better it will be for us all. It's worth the investment in terms of life and livelihood.