MIT AI-Machine Learning Executive Guide: including Deep Learning, Natural Language Processing, Autonomous Cars, Robotic Process Automation
Beyond The Master Algorithm: AI & Machine Learning:
Sense Making for a Non-Deterministic World
Advice to a Management Leadership Executive in the MIT AI-Strategy program
Reproduced below is an updated version of advice shared with a Management Leadership Executive who shared about leading Artificial Intelligence and Machine Learning enterprise adoption and utilization subsequent to participation in one of my sections of the MIT Sloan School of Management & MIT Computer Science & Artificial Intelligence Lab (CSAIL) Program titled AI and Business Strategy for Management and Leadership track industry executives. As evident from the national focus of one of the largest global economies on the book chosen for his internal review and guidance development, The Master Algorithm, he asked me given that the "chosen book is a little older", if I can advice on how to advance beyond The Master Algorithm.
Dear <Management Leadership Executive>,
Thank you for sharing about The Master Algorithm which is an interesting book on the specific 'tribes' of machine learning including Symbolists, Connectionists, Evolutionaries, Bayesians, and, Analogizers.
Given our focus on AI-Strategy, I have introduced through my articles an overarching AI-Strategy framework in particular for advancing beyond 'Prediction' -- which is the focus of the above book among other Computer Science focused Machine Learning books -- to Anticipation of Surprise. The Anticipation of Surprise decision-modeling framework was motivated over two decades ago by the question about what if the past is not an accurate predictor of future as we most often see in the context of Financial Markets, for instance. Prediction based upon historical data may be less useful and likely even dangerous given the risks, uncertainties, and, discontinuities characterizing the business environment in the post-WWW era of exponentially increasing hyper-connectivity and hyper-velocity of global information flows.
I have earlier shared the related focus on the "two worlds of business" emerging at the time of the inception of the World Wide Web. Such emergence of two simultaneous worlds in many organizations, one routine and structured which could be easily automated, and, the other non-routine and unstructured and hence increasingly dependent upon human creativity and ingenuity was observed by other management experts too. At that time, twenty years ago, preceding recent mainstream focus on 'Fluid Intelligence', we distinguished between 'doing the right things' for the latter world of continuous and discontinuous change, and, 'doing things right' as most suited for the former world that was more predictable and certain. The art and science of the organizational strategy as well as the aligned information strategy is in orchestrating the two worlds that exist simultaneously and are also co-evolving together.
Hopefully, the strategic frameworks shared with you in the MIT AI and Business Strategy Program community-of-practice discussions facilitate wholistic understanding of Technological and Process as well as Strategic and Psychological dimensions. Many of our discussions and shared articles further explored their interfaces with Design, Development, and, Deployment of related AI Algorithms and Models. In that context, the specific sections of The Master Algorithm book such as reviewed in the Prologue and key discussions classify many of the machine learning algorithms and related decision modeling frameworks and models across the separate tribes listed earlier.
Within the technical Machine Learning (ML) focus, the tribe with the "backpropagation" seems most related to our Deep Neural Networks (DNN) focus on Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Long Short Term Memory (LSTM) Networks. We have also focused on key algorithms related to the other two tribes with focus on Support Vector Machines (SVM), and, Bayesian Inference. In particular, we have framed our understanding of such AI-ML Algorithms and related Models within a broader socio-technical Strategic view of People, Processes, and, Technology to encompass the Deterministic focus typical of Computer Science books. In that sense, our coverage and discussions provide a much more nuanced view of Human Intelligence, Human Reasoning and Sense Making aligned with our focus on Smart Minds Using Smart Tools Smartly™.
Given our view of People, Processes, and, Technology which is richer in both Strategic and Human aspects, our frameworks of understanding and hopefully resulting 'mindsets' are more in tune with recognizing and managing non-deterministic "wicked" business environments and resulting non-deterministic Futures. In more recent times, specific archetypes of such environments are characterized in terms such as Black Swans and Extreme Events. In a world where many believe Change to be the only constant, our frameworks and mindsets help build the needed self-adaptive capacity and capability for not only surviving in environments characterized by high uncertainty and risk but also thriving in such 'wicked' business environments.
These frameworks, as illustrated in our articles and presentations guide the development and deployment of the "loose-tight" systems as illustrated in one of our earlier presentations. In such 'loose-tight' systems, Sense Making frameworks provide the overarching perspective for increasingly non-deterministic environments while leveraging the relatively deterministic technical AI/ML/DL Information Processing frameworks for more deterministic contexts. The contrast between our Strategic-Human focused approach to People, Processes, and, Technology better tuned to managing uncertainty and change and The Master Algorithm book is illustrated in following representative excerpts from that book. Related comments further clarify our enhanced view that may very well determine the survival and success of the post-WWW Cyber-Crypto-Quantum era digital and data-driven enterprises.
"These seemingly magical technologies work because, at its core, machine learning is about prediction: predicting what we want, the results of our actions, how to achieve our goals, how the world will change... When a new technology is as pervasive and game-changing as machine learning, it’s not wise to let it remain a black box. Opacity opens the door to error and misuse... Not all learning algorithms work the same, and the differences have consequences."
"Hundreds of new learning algorithms are invented every year, but they’re all based on the same few basic ideas. These are what this book is about, and they’re all you really need to know to understand how machine learning is changing the world. Far from esoteric, and quite aside even from their use in computers, they are answers to questions that matter to all of us: How do we learn? Is there a better way? What can we predict? Can we trust what we’ve learned? Rival schools of thought within machine learning have very different answers to these questions."
"The main ones are five in number, and we’ll devote a chapter to each. Symbolists view learning as the inverse of deduction, and take ideas from philosophy, psychology, and logic. Connectionists reverse engineer the brain, and are inspired by neuroscience and physics. Evolutionaries simulate evolution on the computer, and draw on genetics and evolutionary biology. Bayesians believe learning is a form of probabilistic inference, and have their roots in statistics. Analogizers learn by extrapolating from similarity judgments, and are influenced by psychology and mathematical optimization."
From our perspective, 'AI-Strategists' which our group identifies closest with given our shared focus on AI & Business Strategy shall take what makes sense from all the above "tribes" as well as others and apply it to manage Uncertainty and Risk in real world "Messes" (Russell Ackoff) for decision-modeling and decision-making to drive real world performance outcomes.
"Each of the five tribes of machine learning has its own “master algorithm,” a general purpose learner that you can in principle use to discover knowledge from data in any domain. The symbolists’ master algorithm is inverse deduction, the connectionists’ is backpropagation, the evolutionaries’ is genetic programming, the Bayesians’ is Bayesian inference, and the analogizers’ is the support vector machine. In practice, however, each of these algorithms is good for some things but not others. What we really want is a single algorithm combining the key features of all of them: the Master Algorithm. For some this is an unattainable dream, but for many of us in machine learning, it’s what puts a twinkle in our eye and keeps us working late into the night."
"If it exists, the Master Algorithm can derive all knowledge in the world—past, present and future—from data. Inventing it would be one of the greatest advances in the history of science. It would speed up the progress of knowledge across the board, and change the world in ways that we can barely begin to imagine."
"The second goal of this book is thus to enable you to invent the Master Algorithm. You’d think this would require heavy-duty mathematics and severe theoretical work. On the contrary, what it requires is stepping back from the mathematical arcana to see the overarching pattern of learning phenomena; and for this the layman, approaching the forest from a distance, is in some ways better placed than the specialist, already deeply immersed in the study of particular trees. Once we have the conceptual solution, we can fill in the mathematical details; but that is not for this book, and not the most important part."
In our MIT AI Business Strategy Community of Practice "dialog" over the recent few weeks, we have focused on the above approach but taken a broader and more wholistic view of non-deterministic unpredictable contexts and environments in addition to deterministic predictable contexts and environments. In particular, we have emphasized the critical need for going beyond mere data to drive future, such as by relying upon human intuition, imagination, insights, and, creativity. Everything in the world germinates in an idea, an idea has to exist in imagination before it results in any data. Many who are credited for some of the greatest discoveries and innovations in Science and Technology such as Newton and Einstein and others recognized for path-breaking discoveries in fields of Business such as Finance, examples including Fischer Black and Edward Tharp, have recognized the above points as documented by respected scientists and practitioners.
Einstein himself had made the key distinction in following terms: "I am enough of an artist to draw freely upon my imagination. Imagination is more important than knowledge. For knowledge is limited, whereas imagination embraces the entire world, stimulating progress, giving birth to evolution."
"If you’re a machine learning expert, you’re already familiar with much of what the book covers, but you’ll also find in it many fresh ideas, historical nuggets, and useful examples and analogies. Most of all, I hope the book will provide a new perspective on machine learning, and maybe even start you thinking in new directions. Low-hanging fruit is all around us, and it behooves us to pick it, but we also shouldn’t lose sight of the bigger rewards that lie just beyond."
Wishing you all the best in action learning and active practice.