Cognitive Digital Twins at stage 3 reflect the knowledge space of their physical entity (notice that when CDTs have a robot as physical entity they are not reflecting the robot knowledge, they are the robot knowledge) not at a specific point in time but continuously. This requires a continuous update of the CDT through data from the physical entity or more in general data that reflects the change in the physical entity knowledge space. An example might be a CDT created via the KaaS that is receiving information on the articles its physical entity is downloading from the IEEE Xplore, the IEEE conferences, the training courses .attended… Clearly, the information coming from other sources than the physical entity itself (the person) will need to be acknowledged/approved by the physical entity.
Ideally, we could imagine a service in background that has the authorisation to monitor the physical entity and based on the observed behaviour and activity provides updates to the CDT. Notice that in case of personal CDTs (or CDTs resulting from the aggregation of P-CDTs one of the big challenges is to assess what of the knowledge harvested over time is still executable, i,e. is still available to the physical entity (the person). In lay terms, how can a CDT assess what is being forgotten by the associated physical entity? It is a very complex, and “grey”, area but AI based tools can make some reasonable estimate (depending on the latency associated to each knowledge entity -some tend to be forgotten in a short time if not used, are linger for longer time…) also learning from the person behaviour that can provide indication of what is remembered and for how long (memories changes significantly from one person to another and from the situational context).
Interestingly, forgotten knowledge may be refreshed when needed and the CDT may act as a prosthetic. This is also of interest in cases of memory losses (like Alzheimer) and we can expect a growing interest of CDTs in medical domain (in this area “knowledge” has a much broader scope than “knowledge” as defined in the IEEE context).
Cognitive Digital Twins at stage 4 have an expanded knowledge space with respect to the one of their physical entity. There are functions within the CDT to ensure that the CDT:
- is aware of the knowledge space of its physical entity, hence of the difference between its knowledge space and the one of the person;
- has the capability to provide its physical entity seamless access to this extended knowledge space (or to the executable knowledge, meaning that the physical entity can use the impact of the extended knowledge even without being aware of what that knowledge is: like when you use you smartphone calculator app to find a square root what matter is the result of the square root not the algorithm that is being used to calculate the square root).
The expanded knowledge space can be part of the CDT creation or the result of capabilities of the CDT, like use of algorithms (including AI) to access and process data. Additionally, the CDT may have embedded functions that can perform various types of analytics to support the physical entity analyses of the situation and evaluation of alternative reactions.
Based on this we can say that a CDT at stage 4 augments both the knowledge space of its physical entity and augment the capability to use the knowledge (acting like an expert support system).
Notice that, in a way, we already use sort of CDTs at stage 4 every day: we turn to apps on our smartphone to extend our reach to information in a seamless way (just think about the widget you can have on your smartphone presenting you the weather forecast for your location) and we can turn to apps to evaluate a given situation (like knowing where to stop to recharge your electric car taking into account where you are heading, what the traffic situation is, what recharging stations are available). What is missing in this proto-CDTs is a more extended knowledge about “your” knowledge, so that they can be useful in a broader set of everyday situations.
Cognitive Digital Twins at stage 5 become autonomous, in a way independent of their physical entity. However this independence does not diminishes the mirroring of the physical entity knowledge space. In other words, at stage 5 the CDT still has all the properties of a CDT at stage 3.
The independence is playing in two directions:
- acquisition of knowledge
- possibility to act as a knowledge proxy for the physical entity.
The acquisition of knowledge is dynamic and autonomous, that is, the CDT takes actions to expand its knowledge space based on what is relevant for the physical entity and what is available in the cyberspace. As an example, if the physical entity (the person) is working of 5G deployment the CDT keeps monitoring any new products becoming available, all information on issues related to products being used, articles that provide information on the expected evolution in the field… This expanded knowledge will be available to the physical entity in a push-mode: when this knowledge will be needed it will be provided by the CDT. Obviously this requires the CDT to “watch” the physical person to become aware of her needs.
The possibility to act on behalf of the physical person in terms of sharing that person knowledge to others. In this case the person may opt for sharing “only” the knowledge space mirroring the person’s knowledge space or to let the CDT “outperform” the physical person when acting as a proxy. It is already happening. IBM Watson is being used by UBS to let a digital clone of its Chief Economist to meet with multiple clients in parallel, sharing with them the Chief Economist knowledge.
Another possibility is to keep the CDT alive even after its physical entity passed away. In this case other people can keep interacting with the digital copy for long time. This is also a reality today: Replika, using artificial intelligence, offers the possibility to create a digital clone over lasting the physical person.
It is obvious the increasing “augmentation” provided by CDTs as they progress, but this augmentation is just one side of the coin. In the coming posts I will look at the other side.