Published on 12/09/2016 | Technology
Disclaimer: This is an ongoing and partially published work aiming to understand the overall concept of Cognitive Digital Twins (CDT) as the foundation for the next true-smart generation of machines, systems and businesses. It introduces the first definition of CDT and the required system-level cognitive computing architecture. It also summarizes the key categories and relationships of the cognitive digital twins and swarms.
I'm sharing this work here for your own education and hopefully inspiration. Most importantly, as you'll figure out after reading this "long" article, CDTs are a multidiscipline and very complex topic yet very inspiring and promising. Therefore, once you read the article in its entirety, please share it, share your comments and share whatever opinion you might have either online or with me directly.
Being Digital (by Nicholas Negroponte, 1996) and When Things Start to Think (by Neil Gershenfeld, 2000) as well as many R&D efforts done by many teams around the globe over the last decades introduced into the public psyche the potential emergence and the rise of smart machines. Recently, some companies adopted these ideas in their marketing campaign claiming an imminent change to a future where fully integrated and inherently intelligent self-organizing systems, subsystems, and components shall define the next generation of intelligent machines.
Additionally, the massive advances and major breakthroughs occurred in different areas of technology and science gave us a lot of optimism. However, this optimism is faced with the hurdles of removing the dead weight of old technologies and changing our current way of thinking to pave the path for new theories and concepts in the engineering and design of the future systems, which may lead to intelligent machines. An even greater challenge is integrating cognition and evolution capabilities in the design of such machines and systems.
Recently, the discussions about the IoT/IoE, smart machines, and cognitive systems created a lot of motivation, hope, and confusion. Therefore, some companies started different initiatives to discover and understand the possible implications and values of these new technologies to their current and future business. However, without major breakthrough results.
Unfortunately, many of these companies are afraid to delve deeper to be able to envision their future. Instead, they started retrofitting existing machines to designate them as “connected” and attach sensors to collect data in order to call them “smart”. They run workflows on steroids to claim “intelligence” in their analytics. Others are haphazardly collecting and feeding big (volume) data sets to existing software systems and voila – one has “cognitive software systems” listed as an attribute to post their sales numbers forgetting that the harsh reality will soon prove them wrong; the customer experience and opinion.
One reason for the current confusion, perhaps, is our general inability to ask correct questions. Let us step back and pose a few key questions. What machines, devices, and systems could be built with the tools and technologies at hand today and without further inventions? How should we design, build and use them? Do we really want to just connect everything and haphazardly collect large volumes of data? Is it all about the data? How can we enable smart machines to achieve specific tasks while learning to achieve even more and complex tasks over time?
Once we understood and answered those questions, we’ll be able put a plan for relevant innovation and drive effective transformation which will not only help us transform what we’ve to where we want it to be but also, we'll be able to envision and build the right systems, machines and businesses we need in the future.
“We should build smart systems, devices, and machines that can collaborate with us to augment our current physical and cognitive capabilities across and beyond our lifetime and the space we can physically exist in today”
We might agree about this ultimate goal. However, the debate rages on about answers to these and other related questions, which may be cryptic in bio-inspired designs and how nature is evolving to adapt and enhance the physical and cognitive capabilities of creators including humans.
The current breakthroughs in key areas such as “Digital Biology” coupled with Bio-mimicry (https://biomimicry.org/), cybernetics, and deep learning are elements, which when converged may offer completely new dimensions and approaches for real neuromorphic computing and bio-inspired systems and machines that can intelligently sense, learn, reason, act and evolve. This approach may be the Holy Grail.
The current advocacy to advance the principles and pervasive practice of digital twins calls for distributed cognitive and communication capabilities by design where systems can inherit, grow and share intelligence across different sub-systems as well as multiple generations of systems and machines.
Designing, building and managing of such bio-inspired systems and machines may require a new form of hardware-software design, architecture and integration paradigms we haven’t yet encountered. To do so, we’ve to adopt “Cognitive Digital First” way of thinking. This would require us to first design and build cognitive digital twins, which will digitally represent, augment and accompany a new generation of machines (The Physical Twins). Both the digital and physical twins will work in tandem to achieve the excepted tasks and functionalities.
“Cognitive Digital Physical Twins (CDPT) will continue optimizing their cognitive, digital and physical design and capabilities over time based on the data they’ll collect and experience they'll gain, not only based on models and data they we gave them or they inherited”
In this article, while I’ll try to avoid going into technical details and not to give you time-wasting financial predictions, I’ll try to explain my in-progress overall vision, definition, key categories as well as provide an initial reference architecture framework for the Cognitive Digital Twins (CDT) as I see today.
The Noise, confusion and promises of IoT and Digital Transformation
Today, you can’t go to a conference, a meeting or even having a discussion with anybody without having to discuss or hear about the IoT and Digital Transformation. Both are important topics, however as you might have already experienced yourself, most of the discussions end nearly where it started; what is the IoT? Didn’t we already go digital since decades?
This is distracting many people from the real issues and exhausting valuable resources. Therefore, let us get things clearly defined that we can focus on real relevant innovation and challenges.
Today, IoT is being defined as the next generation of digital networking technologies and protocols, which are essentially required to connect a massive number of digital devices and transfer an unlimited amount of data in real time. This is, by all means, not a complete definition.
"The IoT/IoE should also provide the core technologies, platforms, and services required to build the future “Digital Ecology” where smart digital Things can be created, live, smartly act and interact as well as evolve in a manageable and secure way"
And what about Digital Transformation? Recently, companies are asked to digitally transform their business and finally start adopting digital technologies. Really? Didn’t we start digitizing the enterprise tasks, processes and functions many decades ago? And what about the large population of CNC machines and robots adding real values every day and everywhere in many industries?
“The current wave of technology-driven transformation should not be just a technology upgrade (app modernization) or adding more software to what we already have”
Over the past few decades, we went through different waves of technology-driven transformation as follows:
"The required transformation is not only digital, it is a digital cognitive transformation"
To do that, we should understand what does the concept of “Cognitive Digital Enterprise” mean. Cognitive Digital Transformation is way more than adding more analytics tools and gathering more data. Having this as the compass and way of thinking while defining and building the new cognitive enterprise around new smart products and digital services will enable a non-destructive yet efficient transformation to the next generation of smart digitally enabled, represented and augmented business.
The current approach
Some companies got it right and are being able so far to avoid getting stuck and sucked in the wrong and short- sighted terms and discussions about IoT, Digital Transformation, app modernization, Big Data ... etc. However, they’re starting to understand that retrofitting the current physical world as it is today by adding some sensors in and around existing products and haphazardly collecting and analyzing big amount of Data is just a temporary solution that didn’t prove its commercial viability yet and possibly never.
Additionally, it is clear that the current approach of transforming businesses and machines to digitally connected, created major challenges to many organizations without offering a viable proven solution. Some of these challenges are:
- Strategy: What is the right strategy to define and implement a technology-driven new digital business without disrupting the current business?
- Approach: Is the digital twin the right way to go? What does digital twin mean? Is it all about collecting some data?
- Platform: What are the required technology platforms, standards, protocols, and architecture?
- Mist vs. Fog vs. Cloud computing: What about edgeless distributed cognitive computing (EDCC)?
"Distributed Cognitive Architecture will soon end the dilemma of cloud vs. edge vs. mist computing"
- Cognitive capabilities: Do we need to add more data analytics tools or move to the new generation of real cognitive systems that can understand, reason and learn.
- Cyber-physical security: How are we going to proactively identify and prevent the cyber-physical security threats and risks during and after the transformation?
"The current approach of retrofitting machines and tweaking processes is not a feasible and risky approach especially as a long-term solution"
Therefore, many business and technology leaders are looking for a completely new approach to transform their current business and create new ones that they can stay relevant and lead in the era of cognitive business and smart machines.
The Cognitive Digital Twin (CDT): Background
Cyber-Physical Systems (CPS) are heterogeneous blends that combine digital computation, communication and physical world dynamics. They form the foundation of embedded systems (http://leeseshia.org/). The concept of digital twins was first advocated by NASA and became a key engineering concept of its space exploration programs. Advancing the concept of cognitive digital twins in practice must draw heavily from the advances in CPS reengineered using cognitive computing architecture and capabilities.
While cyber-physical systems (CPS) added a lot of values, they produced mainly digitally controlled machines with some embedded software, which can control some functions and collect and analyze little amount of data.
"The current efforts to create “Digital Twins” for existing machines have very limited vision if any. They're limited to a small set of functionalities with the main focus on operational effectivness and predictive maintenance"
"Digital Twins for enterprises are not just a couple of simple analytics cockpits applications"
This is not what a Cognitive Digital Twin of a Physical Twin should be. To get out of the current stagnation and confusing situation, we should start building physical systems, which will be by-design an integral part of intelligent Digital Twins that will accompany and augment them. The physical twin and the digital twin will construct the Cognitive Digital Physical Twin (CDPT).
This mindset and engineering approach are not only about implementing solutions around existing machines or processes to increase productivity or revenue. It is all about building a completely new generation of smart machines, cognitive software systems, as well as digitally connected services, which can be consumed by machines and humans.
Key characteristics and architectural principals of CDTs are:
- CDTs are highly interconnected distributed cognitive systems and in some cases, are very large complex systems of systems (SoSs).
- CDTs exist in the digital space and will evolve over time as their physical Twins evolve.
- CDTs will span physical and digital systems and soon us, the humans.
- They’ll be able to act, interact and collaborate across domains, physical and virtual worlds
- CDTs will continuously evolve to be able to autonomously take contextual decisions and execute more complex tasks in the digital as well as physical worlds (experience-driven behavior)
- CDTs and their physical twins will learn over time to identify, create and provide new services, which their early immature versions couldn’t offer.
- Cognitive digital twins may or may not have physical twins:
Cognitive Digital Twins will have the abilities of physical and digital self-diagnostic and self-healing. For instance,
"CDTs would use techniques such as additive manufacturing or similar future technologies to autonomously design, manufacture and replace some defective parts of its own physical twin."
CDTs will create a new massive economy around semi or fully automated digital smart services defined and offered by the CDTs themselves in collaboration with the physical twin(s) and possibly humans partners.
The early generations of cognitive digital twins will be architected to do no mistakes. Therefore, they’ll have less intelligence and do a lot of mistakes. However, the following generations will be smarter by allowing them to try new things, do mistakes and autonomously learn out of it. Aren’t we so?
“The Cognitive Digital Twin will possess valuable knowledge and experiences gained and optimized over its lifetime. Therefore, it shall not vanish nor be destroyed once the life of the physical twin ended – The power of Mind Transfer and Evolution.”
(Continue reading this article here)
This article was originally posted on LinkedIn.