IEEE Future Directions
Related Articles
Processing data costs close to nothing. If you have a huge amount of data you can extract meaning looking at relations. Think about knowing what medicine you are taking. It provides very little value, surely it does not provide any value to you swallowing the pill every day: you already know it. Now suppose that this information is shared and placed in a set containing related information of thousands of other people swallowing the same pill. Some patterns may emerge, like getting that pill within a few hour of swallowing another specific pill may cause a rash. These sorts of occurrences may be rare and very difficult to spot by a doctor you may visit after the rash. Yet they become straightforward if you (a computer) can make the association. Few years ago,as an example, researchers at Stanford by looking at queries made by millions of people on the Internet were able to discover bad side effects generated by taking two drugs (one for cholesterol and the other for depression).
If you are old enough (nothing to brag about…) you might remember the relational data bases where entities were forming threads based on relations and could be sorted based on those relations. Today, with artificial intelligence and deep learning, we have the possibility of “discovering” relations. This is what can happen if we have digital twins populating the cyberspace of a smart city. An artificial intelligent software can establish relations among digital twins and discover hidden information.
Notice that these relations may be quite sensitive, in terms of privacy, and there should be ways of protecting their visibility. This might be done by implementing query systems governed by the “need to know” rule, and of course there should be a legislation dictating who has the right to express a need (notice that under this paradigm without a need there is no access and the need is traced as well as the information returned to satisfy the need).
The discovered relations among digital twins may be used to create self-organised networks, in this case the artificial intelligence that makes relationships emerge is also the ones that self-organise the network. Notice that the idea is to have these networks as dynamic entities that rearrange themselves based on what the real persons do (this is reflected in their digital twins), hence to have them woven continuously by artificial intelligence is crucial. This is also why I classify these Digital Twins networks as self-organised. By looking at these networks a municipality can promptly identify clusters and their “needs” hence respond to them. Notice the response may also lead to changes in the clusters, different aggregation, splitting and merging. All of this can be simulated in the cyberspace and once the approach leads to satisfactory result it can be implemented and, most important, its effect can be monitored by observing the digital twins and their clustering.
At EIT Digital a number of doctoral students are looking at various aspects of processing data to discover relationships, tackling at the same time privacy and ownership issues. They do this in cooperation with several industries and interact with municipalities. They have been sharing their ideas and have been challenged by industry and institutions representatives starting a dialogue that can foster the digital transformation of our cities.
At the IEEE FDC Initiative on Symbiotic Autonomous Systems (why don’t you join and be part of designing the Future?) Digital Twin is one of the topic under consideration. They are likely to augment us, our mental and physical capabilities enabling exact customisation of robots to our needs.
