There is fierce competition happening in the Graph Technology landscape. In November of last year, I attempted to create a full list [1] of graph technologies. To recap, in this context, "Graph Technology" refers to the use of Graph Databases [2] such as RDF (W3C standard [OWL] - widely adopted and supported by various tools and technologies.) or LPG (Custom-defined schemas and vary from vendor to vendor.) for applications such as Knowledge Graphs [3], this is in contrast to the approach of creating graphs for reporting purposes, such as a network graph (Reporting Technologies).
A whole different technology is involved here. We are talking
about “triples “, “nodes”, “edges” and “hops”. For now, all the vendors have
their own flavour of Query Languages (openCypher, PGQL, GSQL, SQL, G-CORE,
Gremlin, SPARQL, GraphQL, OrientDB, GRAQL), the new GQL World Standard [4] is
on its way.
When looking at the “Gartner - Emerging Tech Impact
Radar: 2023” [5], it can be seen that "Knowledge Graphs" are in the "Critical
Enablers" quadrant with medium to high impact and are expected to become
mainstream within the next 1 to 3 years. This recognition of Knowledge Graphs
as a critical enabler underscores the importance of this technology in enabling
organizations to effectively manage and make sense of their data. To get on the
same page what a “Knowledge Graph” means in this context. It can be used as a
Tool in the Knowledge Management Domain. From an information flow perspective,
using a set of spiderwebs as a metaphor, Knowledge Graphs have the ability to connect
different sets of concepts and relationships within a Graph Database. If
you need interoperability between different technologies, then you will need to
use RDF graph technology. As I
understand it, the new GQL [4] standard will be able to be used on both RDF and
LPG databases, provided that the vendor supports it. Because every LPG
technology has its own flavour, it is important to ensure comprehensive
documentation of development requirements. What will happen if your Star Graph
Database Developer decides to pursue other opportunities and all the tacit,
undocumented knowledge walks out the door? Who will understand what is 'under
the hood'?" (It's ironic that you are creating knowledge tools but not practicing
knowledge management principles.)
There are numerous Use Cases for Knowledge Graphs [2]:
I believe that presently, the Financial and Production
sectors are the ones that benefit the most from graph technologies. Certain Graph
technologies have the ability to track information flow, also known as data
lineage, so you can verify the origin of its source data.
Here are some more examples that demonstrate other
functionality, not limited to:
- Fluree [6] use a graph database, built with
blockchain and semantic graph technology. *
- Ultipa [7] enables microsecond and ultra deep
queries (> 10 hops) on any graph sizes and in real-time. White-box Interpretability. *
- Stardog [8] is using a flexible natural language
interface. *
- TigerGraph’s [9] Machine Learning Workbench. *
- PoolParty’s [10] Graph-Based Text Mining. *
I'm sure this is just the tip of the iceberg when
looking at graph technologies. Other vendors will also have great
functionalities. (Other Graph Technologies [1], you are more than welcome to
add your comments in this post.)
How does the use of Knowledge Graphs relate to
the field of Knowledge Management?
Building on the previous section, which set the
background for 'Practice', let's now explore how to connect different sets of concepts
and relationships, bridging 'Theory' and 'Practice'. Going back to a previous
blog: The tacit knowledge predicament
[11]: “According to Holste et al. (2010): The effective management of tacit
knowledge – the unwritten memory of the firm – is essential to the success of
modern firms. Tacit knowledge is not readily captured or stored by information
technology systems. Increasing investment in information technology will not
translate into better transfer and use or tacit knowledge because individuals
decide whether they will share tacit knowledge and individuals decide whether
they will use tacit knowledge.” [12]
In my opinion, we should focus on the fundamental
principles of Knowledge Management, which involve 'People and Trust'. While
information and technology play a crucial role as tools, the real value of KM
lies in managing knowledge effectively. It is essential to acknowledge that the
term "Knowledge Graphs" can be deceptive and may not accurately
represent its intended meaning. While Knowledge Graphs facilitate advanced
reasoning and analysis beyond the capabilities of traditional databases, a more
accurate term to describe them, in my opinion, would be "Augmented Intelligence
Graphs: A graph-based representation of data that has been enhanced or enriched
in some way (advanced reasoning and analysis) to provide greater value or
insights to users (knowledge management, recommendation systems, and social
network analysis).” In order to connect
people and facilitate tacit knowledge sharing in an informal environment, a
Knowledge Management strategy must include Graph Technologies, such as
Knowledge Graphs. While these tools can assist in Knowledge Management
initiatives, they alone do not drive the field of Knowledge Management. The question that arises is how Knowledge Graphs
(Augmented Intelligence Graphs in my opinion.) can be utilized to encourage
individuals to share their tacit knowledge and remain receptive to the
knowledge shared by others?
From my perspective, the crucial aspect is the ability
to establish relationships within the Knowledge Graph that link employees based
on shared interests in a particular field.
Employees must be allowed to communicate with each
other on an informal basis when they discover a topic that piques their
curiosity within the Knowledge Graph.
Examples:
-
What is this
'Knowledge Graphs' used in the fraud detection department? Maybe we can use it
on our side as well. Who can I speak to?
-
We are starting a
Data Lake project. Who can I speak to about lessons learned with our colleagues
in the UK?"
-
Do we have the
skills available in the organisation to do the project?
-
What does this
mean?
One of the Knowledge Management Policies should be:
"Expect to accept unexpected meeting requests from your colleagues to
share knowledge."
The enabler discussed here is only one aspect of your
Knowledge Management strategy. It's important to strike a balance between a
codification strategy and a personalization strategy, depending on the nature
of your organization's business.
To recap:
-
Do not get confused
with the term “Knowledge Graphs”.
-
Graph technology,
including graph databases and knowledge graphs, can facilitate knowledge
management.
-
See “Knowledge Graphs”
as “Augmented Intelligence Graphs”
-
It is impossible to
codify all tacit knowledge, focus on People and Trust to enable Knowledge sharing.
-
While technology
can support knowledge management, it is unable to manage tacit knowledge on its
own. Its role is primarily to provide a platform that enables people to connect
and share knowledge.
-
Knowledge Graphs
should enable employees to engage in informal interactions based on shared
interests or topics. They should facilitate the connection of information in a
way that makes it easy to find.
-
There needs to be
a balance between a codification strategy and personalization strategy.
[1] https://knowledgemanagement.co.za/blog/kmgraphtech/
[2] https://knowledgemanagement.co.za/blog/graphusecase/
[3] https://knowledgemanagement.co.za/blog/graphusecase/
[4] https://www.gqlstandards.org/existing-languages
[5] https://www.gartner.com/en/doc/emerging-technologies-and-trends-impact-radar-excerpt
[6] https://flur.ee/
[7] https://www.ultipa.com/product/ultipa-graph
[8] https://www.stardog.com/blog/llm-will-accelerate-knowledge-graph-adoption/
[9] https://www.tigergraph.com/ml-workbench/
[10] https://www.poolparty.biz/text-mining-entity-extraction
[11] https://knowledgemanagement.co.za/blog/kmtacitknowldege/
[12] Trust and tacit knowledge sharing and use. Holste,
J Scott; Fields, Dail. Journal of Knowledge Management; Kempston Vol. 14,
Iss. 1, (2010): 135
* At the outset,
we would like to clarify that the purpose of our blogs is to share knowledge
and insights with our readers. Our intention is not to promote any specific
technology or vendor. Any references to products, services or technologies are
made purely for informational purposes and should not be construed as an
endorsement or recommendation.
We strive to ensure that the information presented in
our blogs is accurate, up-to-date and reliable. However, we cannot guarantee
the completeness or accuracy of all the information provided. The opinions and
views expressed in our blogs are those of the author(s) and do not necessarily
reflect the official policies or positions of our organization.
We encourage our readers to do their own research and
exercise their own judgement before making any decisions based on the
information provided in our blogs. We also welcome feedback and comments from
our readers to help us improve the quality and relevance of our content.