I received an early New Year present, actually, it came from Siemens. They had invited me to their Siemens Innovation Day in mid-December 2017. I really appreciated it, yet it took me time to absorb all that was provided, over these past two weeks.
My early present, well actually an idea, came the day before the event. I was included in the Industry Analysts visit to the Siemens Technology Center, at Neuperlach in Munich. We were provided a variety of insights in different presentations and demonstrations, of the technology they are working upon but one stood out for me, being introduced to Knowledge Graphs.
This one ‘thing’ really caught my attention. It was showcased in the technology center, briefly, as part of a broader set of presentations. It immediately struck me as having the potential to be very vital for the connected innovation I see, as our future. These few insights set me off on a new train of thought and I scribbled down some hasty notes while listening to this concept. I then was able to review this a little more after the brief presentation. I then started to research on (Industrial) Knowledge Graphs for the initial depth of understanding I was seeking.
This really excites me in its potential. So let me tell you why
Everything we build upon is based on knowledge. If we retain this knowledge we are re-enforcing our potential for further learning and equally, beginning to question where that existing learning can go. The problem we have, all the time in my case, is our working memory gets layered over with new insights and knowledge. Often we fail to connect up old, already available knowledge with new thinking. We fail to see the connections. We don’t relate or access ‘resident’ knowledge
Why Knowledge Graphs excite me is we can build on Semantic Knowledge, building our Domains of Meaning and fill out the Conceptual Spaces, for adding new meaning and value. We can begin to capture all the knowledge relationships and make those essential connections between data (connecting the dots), deepen and reinforce our learning and then translate them into fresh insights. It can build our ‘collective’ memory, individually and in our communities.
Knowledge Graphs can fill a very big gap.
They can improve our domain knowledge, improve our common sense by linking the relevant context and through a decent labeling of this, we can manage far more complex schemas.
Siemens has partnered, firstly with Metaphacts and built their solutions on the AWS Neptune Engine, offering a query, open language system to build these Knowledge Graphs that works on building relations. Neptune offers knowledge graph that allows you to store information in a graph model and use graph queries, to enable your users to easily navigate highly connected data sets. Neptune supports open source and open standard APIs to allow you to quickly leverage existing information resources, to build your knowledge graphs and host them on a fully managed service.
Using a Resource Description Framework (RDF) a group of companies has been working with this in recent months. Besides Siemens, Astra Zeneca and Thomson Reuters have been looking to explore this. More importantly within Amazon, it is being used for the Amazon Alexa, so this is forming part of their future.
Up to now, knowledge has been kept and used fairly inefficiently, we have tended to keep it in our own way of doing or collecting things, it then becomes rigid in its design and inflexible to radically alter or redesign, without some massive reworking. What if that can change?
What if we can suddenly scale, have availability in different ways and design, be able to query in fast, reliable, easy and open ways? What does that do for knowledge? It opens it up finally, our silos of knowledge become challenged, ready to be built upon and capable of being changed with updated knowledge. Knowledge and insights lead to innovation. Our complete view opens up, it evolves over time and new insights, the more we build and enhance our knowledge, the richer potential we have to build with. Now that sounds tantalizing in what it possibly offers.
So, Siemens is working on Industrial Knowledge Graphs
They start with the premise that human decision-making depends on semantic knowledge for perception, reasoning, and decision-making
They are working to build the world of entities and relations to provide a more intelligible domain model instead of a complex (physical specific) data model. They build the multiple data sources (schema’s) and types of understanding, both structured and unstructured and present these in a formal semantic representation to enable inference (decision potential) and machine processing (decision-making or signaling)
This Machine Intelligence (MI) core can become the engine for realizing more Artificial Intelligence (AI) algorithms, based far more on this constant evolving richer domain knowledge. It reduces uncertainty and prompts greater actions.
So, the industrial knowledge graph sounds really exciting, yet for me, I want to see the Innovation Knowledge Graph as my end result. Now is this a dream or reality? It is different and perhaps more intangible than industrial environments where knowledge is more tangible. You see, I have countless papers, references or visual prompters along with a vast collection of Mind Maps, built for specific innovation scenarios and knowledge capture and use. There is a vast array of insights, knowledge, and advice available within the domain of innovation, can this be captured and translated into an Innovation Knowledge Graph approach? Of course, you can question what is the “right” knowledge? What if I could translate these by capturing and forming better relationships, those can bring into a highly related set of data and knowledge understanding, connecting innovation?
We are all caught up in how digitalization is shaping much of what we do. Yet we all still need to break down isolated pockets of knowledge held in silo’s, we need to access data and knowledge faster and faster, we continue to be inefficient in how we go about this (our workflows) and we often yield far lower quality of outcome than ideal. Can this change? We are still at a very sub-optimal level, in innovation, in knowledge leverage, and in connecting all of our sources of knowledge and data.
What are the contributions of Knowledge Graphs?
Can we extract, generate, link and discover more, by having a system that allows us to connect data, as a graphics-based system, that can present in a highly visual form what we are seeking? Can we get closer to a more complete view, a view of innovation that does, finally, become holistic? Can we build a decent innovation domain ontology that allows expertise to come through, way beyond delivering only pieces of innovation knowledge? Siemens have a clear view it seems of Knowledge Graphs and their contribution to augmented intelligence. Can Innovation Knowledge Graphs help deliver a better relationship we all need in our innovation understanding?
Can we have an interactive application “in our hands” that is highly related in its data, its connections (over time) in relationships, to extract the best expert knowledge from the information, data, and observations, that allow us to focus on the important “sweet spots” of innovation? To actually, constantly innovate, as we build the new, drawn from the knowledge built, on the best known, to form the new understanding, by connecting and forming a relationship, a network of innovation understanding.
What a thought as we enter 2018…….an Innovation Knowledge Graph that takes me closer to a better-connected future for leveraging innovation. Now then, how do I set about it? So I need to first find out more about this and then understand if Knowledge Graphs can be provided for innovation. Intriguing in its possibilities perhaps?