Tag Archives: Machine Learning

Three Types of Actionable Business Analytics Not Called Predictive or Prescriptive

Decades (at least) of business analytics writings have focused on the power, perspicacity, value, and validity in deploying predictive and prescriptive analytics for business forecasting and optimization, respectively. These are primarily forward-looking actionable (proactive) applications.

There are other dimensions of analytics that tend to focus on hindsight for business reporting and causal analysis – these are descriptive and diagnostic analytics, respectively, which are primarily reactive applications, mostly explanatory and investigatory, not necessarily actionable.

In the world of data there are other types of nuanced applications of business analytics that are also actionable – perhaps these are not too different from predictive and prescriptive, but their significance, value, and implementation can be explained and justified differently. Before we dive into these additional types of analytics applications, let us first consider a little pedagogical exercise with two simple evidence-based inferences.

(a) In essentially 100% of cases where an automobile is involved in an accident, the automobile had four wheels on the car prior to the accident.

(b) In 100% of divorce cases, the divorcing couple was married prior to the divorce.

What is the point of those obvious statistical inferences? The point is that the 100% association between the event and the preceding condition has no special predictive or prescriptive power. Hence, prior knowledge of these 100% associations does not offer any actionable value. In statistical terms, the joint probability of event Y and condition X co-occurring, designated P(X,Y), is essentially the probability P(Y) of event Y occurring. The probability of the condition X occurring, P(X), is irrelevant since the existence of the precondition X is implicitly present by default.

Okay, those examples represent two remarkably uninteresting cases. Even when similar sorts of inferences occur in a business context, they have essentially zero value. How do predictive and prescriptive analytics fit into this statistical framework?

Using the same statistical terminology, the conditional probability P(Y|X) (the probability of Y occurring, given the presence of precondition X) is an expression of predictive analytics. By exploring and analyzing the business data, analysts and data scientists can search for and uncover such predictive relationships. This is predictive power discovery. Another way of saying this is: given observed data X, we can predict some outcome Y. Or more simply: given X, find Y.

Similarly (actually, conversely), we can use the conditional probability P(X|Y) (which is the probability that the precondition X exists, given the existence of outcome Y) as an expression of prescriptive analytics. How does that work in practice? By exploring and analyzing business data, analysts and data scientists can search for and uncover the conditions (causal factors) that have led to different outcomes. So, if the business wants to optimize some outcome Y, then data analysts will be tasked with finding the conditions X that must be implemented to achieve that desired outcome. This is prescriptive power discovery. Another way of saying this is: given some desired optimal outcome Y, what conditions X should we put in place. Or more simply: given Y, find X. Note how this simple mathematical expression of prescriptive analytics is exactly the opposite of our previous expression of predictive analytics (given X, find Y).

Here are a few business examples of this type of prescriptive analytics: Which marketing campaign is most efficient and effective (has best ROI) in optimizing sales? Which environmental factors during manufacturing, packaging, or shipping lead to reduced product returns? Which pricing strategies lead to the best business revenue? What equipment maintenance schedule minimizes failures, downtime (mean time to recovery), and overall maintenance costs?

Now that we have described predictive and prescriptive analytics in detail, what is there left? What are the three types of actionable (and valuable) business analytics applications that are not called predictive or prescriptive? They are sentinel, precursor, and cognitive analytics. Let’s define what these are.

Sentinel Analytics – in common usage, the sentinel is the person on the guard station who is charged with watching for significant incoming or emergent activity. In practice, all activity is being observed and a decision is made as to whether any particular activity requires some sort of triage: sounding an alarm, or sending an alert to decision-makers, or doing nothing.
- In the enterprise, sentinel analytics is most timely and beneficial when applied to real-time, dynamic data streams and time-critical decisions. For example, sensors (including internet of things devices and APIs on data networks) can be deployed with logic (analytics, statistical, and/or machine learning algorithms) to monitor and “watch” business systems and processes for emerging patterns, trends, behaviors, unusual operating modes, and anomalies that might be indicators of activities that require business attention, decisions, and/or action.
Precursor Analytics – in common usage, precursors are the early-warning indicators (harbingers, forerunners) of something else more serious or catastrophic that is about to come. We occasionally hear about earthquake precursors (increased levels of radon in groundwater), tidal wave precursors (a deep ocean earthquake), and cyber-attack precursors (phishing incidents). Precursor analytics is related to sentinel analytics. The latter (sentinel) is associated primarily with “watching” the data for interesting patterns that might require action, while precursor analytics is associated primarily with training the business systems to quickly identify those specific “learned” patterns and events that are known to be associated with high-risk events, thus requiring timely attention, intervention, and remediation.
- In these applications, the data science involvement includes both the “learning” of the most significant patterns to alert on and the improvement of their models (logic) to minimize false positives and false negatives. The analytics triage is critical, to avoid alarm fatigue (sending too many unimportant alerts) and to avoid underreporting of important actionable events. One could say that sentinel analytics is more like unsupervised machine learning, while precursor analytics is more like supervised machine learning. That is not a totally clean separation and distinction, but it might help to clarify their different applications of data science.
- The counterexample to the supervised learning explanation of precursor analytics is a “black swan” event – a rare high-impact event that is difficult to predict under normal circumstances – such as the global pandemic, which led to the failure of many predictive models in business. Broken models are definitely disruptive to analytics applications and business operations. Paradoxically, the precursor was actually predictive in a disruptive anti-predictive sort of way, which brings us right back to P(Y|X), or maybe it should be stated as P(“not Y”|X) where X is the black swan event (i.e., the predicted outcome Y from existing models will not occur in this case). As such, the global pandemic serves as a warning (a harbinger of disruption) and consequently as a “training example” to businesses for any future black swans.
Cognitive Analytics – this analytics mindset approach focuses on “surprise” discovery in data, using machine learning and AI to emulate and automate the cognitive abilities of humans. The goal is to discover novel, interesting, unexpected, and potentially valuable signals in the flood of streaming enterprise data. These may not be high-risk discoveries, but they could be high-reward discoveries. How does that resemble human cognitive abilities? Curiosity! Being curious about seeing something “funny” that you didn’t expect, thereby putting a “marker” in the data stream: “Look here! Pay attention! Ask questions about this!”
- Cognitive analytics is basically the opposite of descriptive analytics. In descriptive analytics, the task is to find answers to predetermined business questions (how much, how many, how often, who, where, when), whereas cognitive analytics is tasked with finding the business questions that should be asked. Descriptive: find the right answers in the data. Cognitive: find the right questions in the data. Cognitive analytics can then be viewed as a precursor to diagnostic analytics, which is the investigative stage of analytics that answers the questions raised by cognitive analytics (“Why did this happen?”, “Why are we seeing this pattern in our data?”, “What is the business impact of this trend, anomaly, behavior?”, “What is our next-best action as a result of this?”, “That’s funny! What is that?”).

None of these descriptions of the 3 “new” analytics applications are meant to declare that these are completely distinct and different from the “big 4” analytics applications that we have known for many years (Descriptive, Diagnostic, Predictive, Prescriptive). But the differences between the “big 4” and the “new 3” are in the nuanced business applications of these analytics in the enterprise and in the types of inferences that the data scientists are asked to derive from the business data.

Deploying these analytics in the cloud further expands their accessibility, democratization, enterprise-wide acceptance, broad advocacy, and ultimate business value. Blending automated analytics products (coming from the sentinel, precursor, and cognitive applications) with human-in-the-loop inquisitiveness, curiosity, creativity, out-of-the-box thinking, idea generation, and persistence can transform any organization into a data analytics powerhouse through an analytic culture revolution. This is more imperative than ever, as a global survey of analytics executives has revealed:

“Companies have been working to become more data-driven for many years, with mixed results.”
“Right now, the biggest challenge for organizations working on their data strategy might not have to do with technology at all.”
“Corporate chief data, information, and analytics executives reported that cultural change is the most critical business imperative.”
“Just 26.5% of organizations report having established a data-driven organization.”
“91.9% of executives cite cultural obstacles as the greatest barrier to becoming data driven.”
Reference: https://hbr.org/2022/02/why-becoming-a-data-driven-organization-is-so-hard

Where do organizations get help to overcome these challenges? Microsoft delivers what its clients need to help them grow their top line with cloud-based analytics. Microsoft’s cloud-based analytics products and services propel business insights, innovation, and value from enterprise data, with all of the dimensions of analytics applications brought into the game. Specifically, cloud analytics (accessing and inferencing on multiple diverse business datasets across business units) for a wide variety of enterprise applications can sharpen the workforce’s focus on value and growth, including: forward-looking insights through predictive, sentinel, and precursor analytics; novel recommendations; rich customer engagement; analytic product innovation; resilience through prescriptive analytics; surprise discovery in data, asking the right questions, and exploring the most insightful lines of inquiry through cognitive analytics; and more.

Microsoft Azure Cloud extends ease-of-access analytics to all, delivers increased speed to deployment, provides leading security, compliance, and governance – with price performance for any organization. Whether organizations are seeking scalability in their enterprise data systems, advanced analytics capabilities (including the “big 4” and the “new 3”), real-time analytics (essential value-drivers from streaming data, including IoT, network logs, online customer interactions, supply chain, etc.), and the best in machine learning model-building and deployment services, Microsoft Azure Cloud has you covered. To learn more about it, go to https://azure.microsoft.com/en-us/solutions/cloud-scale-analytics and bring actionable business analytics to higher levels of proficiency and productivity across your organization.

Three Emerging Analytics Products Derived from Value-driven Data Innovation and Insights Discovery in the Enterprise

Leave a reply

I recently saw an informal online survey that asked users which types of data (tabular, text, images, or “other”) are being used in their organization’s analytics applications. This was not a scientific or statistically robust survey, so the results are not necessarily reliable, but they are interesting and provocative. The results showed that (among those surveyed) approximately 90% of enterprise analytics applications are being built on tabular data. The ease with which such structured data can be stored, understood, indexed, searched, accessed, and incorporated into business models could explain this high percentage. A similarly high percentage of tabular data usage among data scientists was mentioned here.

If my explanation above is the correct interpretation of the high percentage, and if the statement refers to successfully deployed applications (i.e., analytics products, in contrast to non-deployed training experiments, demos, and internal validations of the applications), then maybe we would not be surprised if a new survey (not yet conducted) was to reveal that a similar percentage of value-producing enterprise data innovation and analytics/ML/AI applications (hereafter, “analytics products”) are based on on-premises (on-prem) data sources. Why? … because the same productivity benefits mentioned above for tabular data sources (fast and easy data access) would also be applicable in these cases (on-prem data sources). And no one could deny that these benefits would be substantial. What could be faster and easier than on-prem enterprise data sources?

Accompanying the massive growth in sensor data (from ubiquitous IoT devices, including location-based and time-based streaming data), there have emerged some special analytics products that are growing in significance, especially in the context of innovation and insights discovery from on-prem enterprise data sources. These enterprise analytics products are related to traditional predictive and prescriptive analytics, but these emergent products may specifically require low-latency (on-prem) data delivery to support enterprise requirements for timely, low-latency analytics product delivery. These three emergent analytics products are:

(a) Sentinel Analytics – focused on monitoring (“keeping an eye on”) multiple enterprise systems and business processes, as part of an observability strategy for time-critical business insights discovery and value creation from enterprise data sources. For example, sensors can monitor and “watch” systems and processes for emergent trends, patterns, anomalies, behaviors, and early warning signs that require interventions. Monitoring of data sources can include online web usage actions, streaming IT system patterns, system-generated log files, customer behaviors, environmental (ESG) factors, energy usage, supply chain, logistics, social and news trends, and social media sentiment. Observability represents the business strategy behind the monitoring activities. The strategy addresses the “what, when, where, why, and how” questions from business leaders concerning the placement of “sensors” that are used to collect the essential data that power the sentinel analytics product, in order to generate timely insights and thereby enable better data-informed “just in time” business decisions.

(b) Precursor Analytics – the use of AI and machine learning to identify, evaluate, and generate critical early-warning alerts in enterprise systems and business processes, using high-variety data sources to minimize false alarms (i.e., using high-dimensional data feature space to disambiguate events that seem to be similar, but are not). Precursor analytics is related to sentinel analytics. The latter is associated primarily with “watching” the data for interesting patterns, while precursor analytics is associated primarily with training the business systems to quickly identify those specific patterns and events that could be associated with high-risk events, thus requiring timely attention, intervention, and remediation. One could say that sentinel analytics is more like unsupervised machine learning, while precursor analytics is more like supervised machine learning. That is not a totally clear separation and distinction, but it might help to clarify their different applications of data science. Data scientists work with business users to define and learn the rules by which precursor analytics models produce high-accuracy early warnings. For example, an exploration of historical data may reveal that an increase in customer satisfaction (or dissatisfaction) with one particular product is correlated with some other satisfaction (or dissatisfaction) metric downstream at a later date. Consequently, based on this learning, deploying a precursor analytics product to detect the initial trigger event early can thus enable a timely response to the situation, which can produce a positive business outcome and prevent an otherwise certain negative outcome.

(c) Cognitive Analytics – focused on “surprise” discovery in diverse data streams across numerous enterprise systems and business processes, using machine learning and data science to emulate and automate the curiosity and cognitive abilities of humans – enabling the discovery of novel, interesting, unexpected, and potentially business-relevant signals across all enterprise data streams. These may not be high risk. They might actually be high-reward discoveries. For example, in one company, an employee noticed that it was the customer’s birthday during their interaction and offered a small gift to the customer at that moment—a gift that was pre-authorized by upper management because they understood that their employees are customer-facing and they anticipated that their employees would need to have the authority to take such customer-pleasing actions “in the moment”. The outcome was very positive indeed, as this customer reported the delightful experience on their social media account, thereby spreading positive sentiment about the business to a wide audience. Instead of relying on employees to catch all surprises in the data streams, the enterprise analytics applications can be trained to automatically watch for, identify, and act on these surprises. In the customer birthday example, the cognitive analytics product can be set up for automated detection and response, which can occur without the employee in the loop at all, such as in a customer’s online shopping experience or in a chat with the customer call center bot.

These three analytics products are derived from business value-driven data innovation and insights discovery in the enterprise. Investigating and deploying these are a worthy strategic move for any organization that is swimming in a sea (or lake or ocean) of on-prem enterprise data sources.

In closing, let us look at some non-enterprise examples of these three types of analytics:

Sentinel – the sentinel on the guard station at a military post is charged with watching for incoming activity. They are assigned this duty just in case something occurs during the night or when everyone else is busy with other operational things. That “something” might be an enemy approaching or a wild bear in the forest. In either case, keeping an eye on the situation is critical for the success of the operation. Another example of a sentinel is a marked increase in the volatility of stock market prices, indicating that there may be a lot of FUD (fear, uncertainty, and doubt) in the market that could lead to wild swings or downturns. In fact, anytime that any streaming data monitoring metric shows higher than usual volatility, this may be an indicator that the monitored thing requires some attention, an investigation, and possibly an intervention.
Precursor – prior to large earthquakes, it has been found that increased levels of radon are detected in soil, in groundwater, and even in the air in people’s home basements. This precursor is presumed to be caused by the radon being released from cavities within the Earth’s crust as the crust is being strained prior to the sudden slippage (the earthquake). Earthquakes themselves can be precursors to serious events – specifically, a large earthquake detected at the bottom of the ocean can produce a massive tidal wave, that can travel across the ocean and have drastic consequences on distant shores. In some cases, the precursor can occur sufficiently in advance of the tidal wave’s predicted arrival at inhabited shores, thereby enabling early warnings to be broadcasted. In both of these cases, the precursor (radon release or ocean-based earthquake) is not the biggest problem, though they may be seen as sentinels of an on-going event, but the precursor is an early warning sign of a potentially bigger catastrophe that’s coming (a major land-based earthquake or a tidal wave hitting major population centers along coastlines, respectively).
Cognitive – a cognitive person walking into an intense group meeting (perhaps a family or board meeting) can probably tell the mood of the room fairly quickly. The signals are there, though mostly contextual, thus probably missed by a cognitively impaired person. A cognitive person is curious about odd things that they see and hear—things or circumstances or behaviors that seem out of context, unusual, and surprising. The thing itself (or the data about the thing) may not be surprising (though it could be), but the context (the “metadata”, which is “other data about the primary data”) provides a signal that something needs attention here. Perhaps the simplest expression of being cognitive in this data-drenched world comes from a quote attributed to famous science writer Isaac Asimov: “The most exciting phrase to hear in science, the one that heralds new discoveries, is not ‘Eureka!’ (I found it!) but ‘That’s funny…‘.”

The cognitive enterprise versus the cognitively impaired enterprise – which of these would your organization prefer to be? Get moving now with sentinel, precursor, and cognitive analytics through data innovation and insights discovery with your on-prem enterprise data sources.

Read more about analytics innovation from on-prem enterprise data sources in this 3-part blog series:

Source: https://www.azquotes.com/author/605-Isaac_Asimov

Top 9 Considerations for Enterprise AI

Leave a reply

Artificial intelligence (AI) is top of mind for executives, business leaders, investors, and most workplace employees everywhere. The impacts are expected to be large, deep, and wide across the enterprise, to have both short-term and long-term effects, to have significant potential to be a force both for good and for bad, and to be a continuing concern for all conscientious workers. In confronting these winds of change, enterprise leaders are faced with many new questions, decisions, and requirements – including the big question: are these winds of change helping us to move our organization forward (tailwinds) or are they sources of friction in our organization (headwinds)?

The current AI atmosphere in enterprises reminds us of the internet’s first big entrance into enterprises nearly three decades ago. I’m not referring to the early days of email and Usenet newsgroups, but the tidal wave of Web and e-Commerce applications that burst onto the business scene in the mid-to-late 1990’s. While those technologies brought much value to the enterprise, they also brought an avalanche of IT security concerns into the C-suite, leading to more authoritative roles for the CIO and the CISO. The fraction of enterprise budgets assigned to these IT functions (especially cybersecurity) suddenly and dramatically increased. That had and continues to have a very big and long-lasting impact.

The Web/e-Commerce tidal wave also brought a lot of hype and FOMO, which ultimately led to the Internet bubble burst (the dot-com crash) in the early 2000’s. AI, particularly the new wave of generative AI applications, has the potential to repeat this story, potentially unleashing a wave of similar patterns in the enterprise. Are we heading for another round of hype / high hopes / exhilaration / FOMO / crash and burn with AI? I hope not.

I would like to believe that a sound, rational, well justified, and strategic introduction of the new AI technologies (including ChatGPT and other generative AI applications) into enterprises can offer a better balance on the fast slopes of technological change (i.e., protecting enterprise leaders from getting out too far over their skis). In our earlier article, we discussed “AI Readiness is Not an Option.” In this article here, we offer some considerations for enterprise AI to add to those strategic conversations. Specifically, we look at considerations from the perspective of the fuel for enterprise AI applications: the algorithms, the data, and the enterprise AI infrastructure. Here is my list:

[continue reading the full article here]

https://blog.purestorage.com/solutions/ai-breakthrough-awards-best-ai-solution-airi-s/

AI Readiness is Not an Option

Leave a reply

This year, artificial intelligence (AI) has become a major conversation centerpiece at home, in the park, at the gym, at work, everywhere. This is not entirely due to or related to ChatGPT and LLMs (large language models), though those have been the main drivers. The AI conversations, especially in technical circles, have focused intensively on generative AI, the creation of written content, images, videos, marketing copy, software code, speeches, and countless other things. For a short introduction to generative AI, see my article “Generative AI – Chapter 1, Page 1”.

While there has been huge public interest in generative AI (specifically, ChatGPT) by individuals, there has been a transformative impact on organizations everywhere, both in strategy conversations and tactical deployments. Businesses and others are seeking to leverage generative AI to increase productivity (efficiencies and effectiveness) in nearly all aspects of their enterprise.

To support essential enterprise AI strategy conversations, here are 12 key points for organizations to consider within the context of “AI readiness is not an option, but an imperative”:

[continue reading the full article here]

Built for AI – https://purefla.sh/41oS2Dp

Generative AI – Chapter 1, Page 1

Leave a reply

Anyone who has been watching the AI space this year, even peripherally, will have noticed the flaming hot story of the year—ChatGPT and related chatbot applications. These AI applications are essentially deep machine learning models that are trained on hundreds of gigabytes of text and that can provide detailed, grammatically correct, and “mostly accurate” text responses to user inputs (questions, requests, or queries, which are called prompts). Specifically, these are LLMs—large language models. It is imperative, not an option, for organizations (and for most individuals) to be aware of what is going on here—not only because it is all over the news, but because it could affect your future self.

When I said “mostly accurate,” I meant that sometimes the ChatGPT responses go way off target—people refer to these as “hallucinations,” which is basically a reflection of the statistical basis of the models (see below)—the application will generate some plausible-sounding, grammatically correct statements that are complete falsehoods, such as “Leonardo da Vinci painted the Mona Lisa in 1815” (which is a real example of an observed ChatGPT hallucination).

I tested ChatGPT with my own account, and I was impressed with the results. I prompted it with various requests, including: Write a short story on a specific topic, provide a layperson’s explanations of some complex deep machine learning concepts, create a lesson plan to learn a tough subject, create an outline for a blog on a particular topic (no, not this one), and provide some financial advice on particular investments (no, it did not provide specific advice, but it did offer warnings like NFA “Not Financial Advice” and DYOR “Do Your Own Research”). You can find my results on my Medium blog site.

LLMs are so responsive and grammatically correct (even over many paragraphs of text) that some people worry that it is sentient. Guess what? It isn’t. It is merely a very large statistical model that provides the most likely sequence of words in response to a prompt. It is effectively a galaxy-sized statistically rich version of text autocomplete on your smartphone’s text messaging app, which already delivers some highly probable guesses for the missing words in a text message like this one: “Due to a client deadline, I will be working late at the ____ this ____, so I will be home late for ____.” LLMs can respond to much more complex (but well-posed) prompts, such as lesson plans for education, content for a business presentation, code for a software task, workflow steps for an IT project, and much more.

In order to help people to create well-posed prompts, the new discipline of prompt engineering has arisen. It’s not hard to find many online guides to prompt engineering, including guides for very specific industries, business tasks, workplace applications, and context-dependent scenarios. You don’t need prompt engineering to find those guides—a simple web search should do the trick. And guess what? When web search engines were first created, it took a while for us to learn how to submit well-posed keyword searches. That scenario is being played out again with ChatGPT and prompt engineering, but now our queries are aimed at a much more language-based, AI-powered, statistically rich application. If you understand Bayes’ Theorem and Bayesian statistics, then you will understand me when I say that we are talking here about an enormously more enriched set of priors, likelihoods, and evidence to feed the LLMs—so, it should not be surprising that the posteriors are shockingly good for large text outputs (most of the time).

LLMs are a subset of the deep learning field of natural language processing (NLP), which includes natural language understanding (NLU) and natural language generation (NLG). Think of chatbots and you get the idea, just expanded to a much, much larger domain of AI-based conversation.

Computer vision (CV) is another subset of deep learning, specifically aimed at object/pattern detection, recognition, and classification in images (including still images and video sequences). ChatGPT and LLMs are examples of generative AI using NLP for text generation. Stable Diffusion, Midjourney, and Dall-E are examples of generative AI using CV for image generation. Oh, by the way, I asked the generative AI at Stable Diffusion to create some images to go with my short story (which you can find on my Medium blog).

Beyond the individual examples of generative AI (and its components, ChatGPT, Stable Diffusion, etc.) that we can all experiment with, the applications in the enterprise can be tremendously impactful and transformative for organizations and the future of work. Those next chapters in the story are being written right now.

Continue reading about Enterprise AI in these posts:

Source: https://marketoonist.com/2023/01/ai-tidal-wave.html

Business Strategies for Deploying Disruptive Tech: Generative AI and ChatGPT

Leave a reply

Generative AI is the biggest and hottest trend in AI (Artificial Intelligence) at the start of 2023. While generative AI has been around for several years, the arrival of ChatGPT (a conversational AI tool for all business occasions, built and trained from large language models) has been like a brilliant torch brought into a dark room, illuminating many previously unseen opportunities.

Source: https://www.prnewswire.com/news-releases/the-future-of-chatgpt-and-generative-ai-in-the-enterprise-according-to-info-tech-research-group-301705286.html

Every business wants to get on board with ChatGPT, to implement it, operationalize it, and capitalize on it. It is important to realize that the usual “hype cycle” rules prevail in such cases as this. First, don’t do something just because everyone else is doing it – there needs to be a valid business reason for your organization to be doing it, at the very least because you will need to explain it objectively to your stakeholders (employees, investors, clients). Second, doing something new (especially something “big” and disruptive) must align with your business objectives – otherwise, you may be steering your business into deep uncharted waters that you haven’t the resources and talent to navigate. Third, any commitment to a disruptive technology (including data-intensive and AI implementations) must start with a business strategy.

I suggest that the simplest business strategy starts with answering three basic questions: What? So what? Now what? That is: (1) What is it you want to do and where does it fit within the context of your organization? (2) Why should your organization be doing it and why should your people commit to it? (3) How do we get started, when, who will be involved, and what are the targeted benefits, results, outcomes, and consequences (including risks)? In short, you must be willing and able to answer the six WWWWWH questions (Who? What? When? Where? Why? and How?).

Another strategy perspective on technology-induced business disruption (including generative AI and ChatGPT deployments) is to consider the three F’s that affect (and can potentially derail) such projects. Those F’s are: Fragility, Friction, and FUD (Fear, Uncertainty, Doubt).

Fragility occurs when a built system is easily “broken” when some component is changed. These changes may include requirements drift, data drift, model drift, or concept drift. The first one (requirements drift) is a challenge in any development project (when the desired outcomes are changed, sometimes without notifying the development team), but the latter three are more apropos to data-intensive product development activities (which certainly describes AI projects). A system should be sufficiently agile and modular such that changes can be made with as little impact to the overall system design and operations as possible, thus keeping the project off the pathway to failure. Since ChatGPT is built from large language models that are trained against massive data sets (mostly business documents, internal text repositories, and similar resources) within your organization, consequently attention must be given to the stability, accessibility, and reliability of those resources.

Friction occurs when there is resistance to change or to success somewhere in the project lifecycle or management chain. This can be overcome with small victories (MVP minimum viable products, or MLP minimum lovable products) and with instilling (i.e., encouraging and rewarding) a culture of experimentation across the organization. When people are encouraged to experiment, where small failures are acceptable (i.e., there can be objective assessments of failure, lessons learned, and subsequent improvements), then friction can be minimized, failure can be alleviated, and innovation can flourish. A business-disruptive ChatGPT implementation definitely fits into this category: focus first on the MVP or MLP.

FUD occurs when there is too much hype and “management speak” in the discussions. FUD can open a pathway to failure wherever there is: (a) Fear that the organization’s data-intensive, machine learning, AI, and ChatGPT activities are driven by FOMO (fear of missing out, sparked by concerns that your competitors are outpacing your business); (b) Uncertainty in what the AI / ChatGPT advocates are talking about (a “Data Literacy” or “AI Literacy” challenge); or (c) Doubt that there is real value in the disruptive technology activities (due to a lack of quick-win MVP or MLP examples).

I have developed a few rules to help drive quick wins and facilitate success in data-intensive and AI (e.g., Generative AI and ChatGPT) deployments. These rules are not necessarily “Rocket Science” (despite the name of this blog site), but they are common business sense for most business-disruptive technology implementations in enterprises. Most of these rules focus on the data, since data is ultimately the fuel, the input, the objective evidence, and the source of informative signals that are fed into all data science, analytics, machine learning, and AI models.

Here are my 10 rules (i.e., Business Strategies for Deploying Disruptive Data-Intensive, AI, and ChatGPT Implementations):

Honor business value above all other goals.
Begin with the end in mind: goal-oriented, mission-focused, and outcomes-driven, while being data-informed and technology-enabled.
Think strategically, but act tactically: think big, start small, learn fast.
Know thy data: understand what it is (formats, types, sampling, who, what, when, where, why), encourage the use of data across the enterprise, and enrich your datasets with searchable (semantic and content-based) metadata (labels, annotations, tags). The latter is essential for AI implementations.
Love thy data: data are never perfect, but all the data may produce value, though not immediately. Clean it, annotate it, catalog it, and bring it into the data family (connect the dots and see what happens). For example, outliers are often dismissed as random fluctuations in data, but they may be signaling at least one of these three different types of discovery: (a) data quality problems, associated with errors in the data measurement and capture processes; (b) data processing problems, associated with errors in the data pipeline and transformation processes; or (c) surprise discovery, associated with real previously unseen novel events, behaviors, or entities arising in your data stream.
Do not covet thy data’s correlations: a random six-sigma event is one-in-a-million. So, if you have 1 trillion data points (e.g., a Terabyte of data), then there may be one million such “random events” that will tempt any decision-maker into ascribing too much significance to this natural randomness.
Validation is a virtue, but generalization is vital: a model may work well once, but not on the next batch of data. We must monitor for overfitting (fitting the natural variance in the data), underfitting (bias), data drift, and model drift. Over-specifying and over-engineering a model for a data-intensive implementation will likely not be applicable to previously unseen data or for new circumstances in which the model will be deployed. A lack of generalization is a big source of fragility and dilutes the business value of the effort.
Honor thy data-intensive technology’s “easy buttons” that enable data-to-discovery (D2D), data-to-“informed decision” (D2ID), data-to-“next best action” (D2NBA), and data-to-value (D2V). These “easy buttons” are: Pattern Detection (D2D), Pattern Recognition (D2ID), Pattern Exploration (D2NBA), and Pattern Exploitation (D2V).
Remember to Keep it Simple and Smart (the “KISS” principle). Create a library of composable, reusable building blocks and atomic business logic components for integration within various generative AI implementations: microservices, APIs, cloud-based functions-as-a-service (FaaS), and flexible user interfaces. (Suggestion: take a look at MACH architecture.)
Keep it agile, with short design, develop, test, release, and feedback cycles: keep it lean, and build on incremental changes. Test early and often. Expect continuous improvement. Encourage and reward a Culture of Experimentation that learns from failure, such as “Test, or get fired!”

Finally, I offer a very similar (shorter and slightly different) set of Business Strategies for Deploying Disruptive Data-Intensive, AI, and ChatGPT Implementations, from the article “The breakthrough that is ChatGPT: How much does it cost to build?“. Here is the list from that article’s “C-Suite’s Guide to Developing a Successful AI Chatbot”:

Define the business requirements.
Conduct market research.
Choose the right development partner.
Develop a minimum viable product (MVP).
Test and refine the chatbot.
Launch the chatbot.

Graphic source: https://appinventiv.com/blog/cost-to-develop-an-app-like-chatgpt/

My top learning and pondering moments at Splunk .conf22

Leave a reply

I recently attended the Splunk .conf22 conference. While the event was live in-person in Las Vegas, I attended virtually from my home office. Consequently I missed the incredible in-person experience of the brilliant speakers on the main stage, the technodazzle of 100’s of exhibitors’ offerings in the exhibit arena, and the smooth hip hop sounds from the special guest entertainer — guess who?

What I missed in-person was more than compensated for by the incredible online presentations by Splunk leaders, developers, and customers. If you have ever attended a major expo at one of the major Vegas hotels, you know that there is a lot of walking between different sessions — literally, miles of walking per day. That’s good for you, but it often means that you don’t attend all of the sessions that you would like because of the requisite rushing from venue to venue. None of that was necessary on the Splunk .conf22 virtual conference platform. I was able to see a lot, learn a lot, be impressed a lot, and ponder a lot about all of the wonderful features, functionalities, and future plans for the Splunk platform.

One of the first major attractions for me to attend this event is found in the primary descriptor of the Splunk Platform — it is appropriately called the Splunk Observability Cloud, which includes an impressive suite of Observability and Monitoring products and services. I have written and spoken frequently and passionately about Observability in the past couple of years. For example, I wrote this in 2021:

“Observability emerged as one of the hottest and (for me) most exciting developments of the year. Do not confuse observability with monitoring (specifically, with IT monitoring). The key difference is this: monitoring is what you do, and observability is why you do it. Observability is a business strategy: what you monitor, why you monitor it, what you intend to learn from it, how it will be used, and how it will contribute to business objectives and mission success. But the power, value, and imperative of observability does not stop there. Observability meets AI – it is part of the complete AIOps package: ‘keeping an eye on the AI.’ Observability delivers actionable insights, context-enriched data sets, early warning alert generation, root cause visibility, active performance monitoring, predictive and prescriptive incident management, real-time operational deviation detection (6-Sigma never had it so good!), tight coupling of cyber-physical systems, digital twinning of almost anything in the enterprise, and more. And the goodness doesn’t stop there.”
Continue reading my thoughts on Observability at http://rocketdatascience.org/?p=1589

The dominant references everywhere to Observability was just the start of awesome brain food offered at Splunk’s .conf22 event. Here is a list of my top moments, learnings, and musings from this year’s Splunk .conf:

Observability for Unified Security with AI (Artificial Intelligence) and Machine Learning on the Splunk platform empowers enterprises to operationalize data for use-case-specific functionality across shared datasets. (Reference)
The latest updates to the Splunk platform address the complexities of multi-cloud and hybrid environments, enabling cybersecurity and network big data functions (e.g., log analytics and anomaly detection) across distributed data sources and diverse enterprise IT infrastructure resources. (Reference)
Splunk Enterprise 9.0 is here, now! Explore and test-drive it (with a free trial) here.
The new Splunk Enterprise 9.0 release enables DevSecOps users to gain more insights from Observability data with Federated Search, with the ability to correlate ops with security alerts, and with Edge Management, all in one platform. (Reference)
Security information and event management (SIEM) on the Splunk platform is enhanced with end-to-end visibility and platform extensibility, with machine learning and automation (AIOps), with risk-based alerting, and with Federated Search (i.e., Observability on-demand). (Reference)
Customer success story: As a customer-obsessed bank with ultra-rapid growth, Nubank turned to Splunk to optimize data flows, analytics applications, customer support functions, and insights-obsessed IT monitoring. (Reference)
The key characteristics of the Splunk Observability Cloud are Resilience, Security, Scalability, and EXTENSIBILITY. The latter specifically refers to the ease in which developers can extend Splunk’s capabilities to other apps, applying their AIOps and DevSecOps best practices and principles! Developers can start here.
The Splunk Observability Cloud has many functions for data-intensive IT, Security, and Network operations, including Anomaly Detection Service, Federated Search, Synthetic Monitoring, Incident Intelligence, and much more. Synthetic monitoring is essentially digital twinning of your network and IT environment, providing insights through simulated risks, attacks, and anomalies via predictive and prescriptive modeling. [Reference]
Splunk Observability Cloud’s Federated Search capability activates search and analytics regardless of where your data lives — on-site, in the cloud, or from a third party. (Reference)
The new release of the Splunk Data Manager provides a simple, modern, automated experience of data ingest for Splunk Cloud admins, which reduces the time it takes to configure data collection (from hours/days to minutes). (Reference)
Splunk works on data, data, data, but the focus is always on customer, customer, customer — because delivering best outcomes for customers is job #1. Explore Splunk’s amazing Partner ecosystem (Partnerverse) and the impressive catalog of partners’ solutions here.
Splunk .conf22 Invites Organizations to Unlock Innovation With Data.

In summary, here is my list of key words and topics that illustrate the diverse capabilities and value-packed features of the Splunk Observability Cloud Platform that I learned about at the .conf22 event:

– Anomaly Detection Assistant
– Risk-based Alerting (powered by AI and Machine Learning scoring algorithms)
– Federated Search (Observability on-demand)
– End-to-End Visibility
– Platform Extensibility
– Massive(!) Scalability of the Splunk Observability Cloud (to billions of transactions per day)
– Insights-obsessed Monitoring (“We don’t need more information. We need more insights.”)
– APIs in Action (to Turn Data into Doing™)
– Splunk Incident Intelligence
– Synthetic Monitoring (Digital Twin of Network/IT infrastructure)
– Splunk Data Manager
– The Splunk Partner Universe (Partnerverse)

My closing thought — Cybersecurity is basically Data Analytics: detection, prediction, prescription, and optimizing for unpredictability. This is what Splunk lives for!

Follow me on LinkedIn here and on Twitter at @KirkDBorne.

Disclaimer: I was compensated as an independent freelance media influencer for my participation at the conference and for this article. The opinions expressed here are entirely my own and do not represent those of Splunk or of any Splunk partners. Any misrepresentations of the products and services mentioned in my statements are entirely my own responsibility. Nothing here should be construed as an offer to sell or as financial advice of any kind. My comments are entirely of a technical nature, focused on the technical capabilities of the items mentioned in the article.

Splunk Cloud Platform Dashboard.
Source: https://www.splunk.com/en_us/products/splunk-cloud-platform.html

Data Science Blogs-R-Us

Leave a reply

[UPDATED December 31, 2022]

I have written articles in many places. I will be collecting links to those sources here. The list is not complete and will be constantly evolving. There are some older blogs that I will be including in the list below as I remember them and find them. Also included are some interviews in which I provided detailed answers to a variety of questions.

In 2019, I was listed as the #1 Top Data Science Blogger to Follow on Twitter.

And then there’s this — not a blog, but a link to my 2013 TedX talk: “Big Data, Small World.” (Many more videos of my talks are available online. That list will be compiled in another place soon.)

Glossaries of Data Science Terminology

Leave a reply

Here is a compilation of glossaries of terminology used in data science, big data analytics, machine learning, AI, and related fields:

A tag cloud of data science and machine learning terminology

Data Science Training Opportunities

Leave a reply

A few years ago, I generated a list of places to receive data science training. That list has become a bit stale. So, I have updated the list, adding some new opportunities, keeping many of the previous ones, and removing the obsolete ones.

Also, here is a thorough, informative, and interesting article that outlines the critical skills needed in order to be a good data scientist: https://www.toptal.com/data-science#hiring-guide

Here are 30 training opportunities that I encourage you to explore:

Follow Kirk Borne on Twitter @KirkDBorne

Field Guide to Data Science — Learn the what, why, and how of Data Science and Machine Learning here.

Rocket-Powered Data Science

Data Reflections by Dr. Kirk Borne @KirkDBorne

Tag Archives: Machine Learning

Three Types of Actionable Business Analytics Not Called Predictive or Prescriptive

Three Emerging Analytics Products Derived from Value-driven Data Innovation and Insights Discovery in the Enterprise

Top 9 Considerations for Enterprise AI

AI Readiness is Not an Option

Generative AI – Chapter 1, Page 1

Business Strategies for Deploying Disruptive Tech: Generative AI and ChatGPT

My top learning and pondering moments at Splunk .conf22

Data Science Blogs-R-Us

Glossaries of Data Science Terminology

Data Science Training Opportunities