7. Knowledge platforms
I closed the last post touching on “knowledge” and pointing out that our smartphones, our digital twins, might become knowledge intermediators in our access to the knowledge on the web.
This ties in nicely with the last category of platforms I would like to discuss: knowledge platforms.
If you google for knowledge platform you will get plenty of pointers to companies and services delivering various forms of knowledge, each one specialised in a given sector, like agriculture, high school math, do-it-yourself advice and so on.
What I am referring to, is something different: a platform that supports the creation and sharing of knowledge at semantic level, taking into account the needs and capability of the person or entity requiring such knowledge.
Today we have plenty of platform dealing at a “syntactic” level, easing the exchange of documents (in various forms). The various education platforms are an example, like Coursera and EdX. IEEE Xplore is another example of a platform for sharing documents. I prefer to call these platforms rather than data bases because they are providing a sort of rubber stamp on the documents being shared (in the case of Xplore all documents have been peer-reviewed, in case of Coursera and EdX the courses have been checked for quality and so on).
Current education platforms are not that different from general content platforms, like Netflix or YouTube. These latter also have a (sometimes minimally) curated content, and sometimes that content is also of educational nature, designed to share knowledge. However, there is little, if any, consideration to the user capability to acquire the knowledge in the way it is presented.
On YouTube I can watch a clip in Chinese, and the transfer of knowledge, given my understanding of Chinese, will be zero. Yet, YouTube will duly show me that Chinese clip. Similarly, I can ask to follow a course on Coursera without the pre-required knowledge that would allow me to understand it, hence acquire the knowledge being presented. I can log on Xplore and read a paper that although written in English looks Arabic to me. Again: net transfer of knowledge equals 0.
There are now studies on Cognitive Digital Twins that may result in the creation of what I am calling knowledge platforms.
At IEEE FDC we are initiating an ambitious project to develop a knowledge platform that can transform the huge content base of Xplore into Knowledge-as-a-Service. This is done by creating cognitive digital twins of the users of Xplore (using an opt-in approach) and having those digital twins interacting with the KaaS to transfer knowledge on demand as well as to be notified on the availability of knowledge that may be needed by that person, given her current activities or the ones expected in the near future. As an example, a cognitive digital twin of a person who took a WCET (Wireless Certification) this year will be notified of the evolution in wireless technology, systems, applications that suggests the need for updating the knowledge of that person. The cognitive digital twin, being aware of the current –and short terms- needs of the person will interact with Xplore, and with the IEEE education courses portfolio, to design a roadmap of knowledge transfer that may fit that person (also taking into account her time availability).
The project is staged in three phases, one per each of the coming years:
- 2020 – delivery of a web interface to create the person’s cognitive digital twin
- 2021 – delivery of tools to assess needs and extract/render knowledge
- 2022 – delivery of a knowledge map, dynamically tuned based on the changing user needs and on the evolving knowledge space that can be used to identify knowledge gaps and actions to fill them
The program is targeting first the needs of personal knowledge and then the one of enterprise, institution knowledge with the ambition of providing knowledge services of various types, from course to consultancy and resources (people, companies, AI tools).
Something that gets close to this concept of knowledge platform is the one of UnanimousAI. This is a company, and a service, that share knowledge in the medical domain. It assumes that those connecting to it are medical doctors, hence it presents knowledge in a way that is customised for them. The knowledge base is created in two ways:
– by roaming the web to retrieve the latest information made available in the medical field (accessing trusted magazines, journals and data banks in the medical area) and extracting knowledge using artificial intelligence software;
– by growing the knowledge base through interaction with medical professionals that every day present cases in their practice, asking for reference to other doctors who may have faced similar cases.
The access to the knowledge base is mediated by natural language based interfaces and the more interactions take place the more UnanimousAI grow its knowledge through Machine Learning.