By Ahren Lehnert and Bob Kasenchak
The yearly KMWorld 2019 Conference, and its subsidiary conferences Enterprise Search & Discovery, Office 365 Symposium, Taxonomy Boot Camp, Text Analytics Forum, and first-time Complexity in Human Systems Symposium, took place November 4-7 in Washington, D.C. The Synaptica team was there representing our products, giving presentations, and attending as many sessions as possible.
While the conferences are too expansive and overlapping to attend as many sessions as we would like, we do have some general observations.
Our big takeaways are that knowledge management, discovery, and delivery are maturing rapidly with the rise and improvement of new technologies. Especially notable is that knowledge graphs were discussed at sessions in every conference track.
Taxonomy Boot Camp is enjoying growth; chair Stephanie Lemieux noted that attendance has continued to increase each of the past three years. In addition to the annual appearance of the usual suspects (consultants, software companies, etc.) in the world of taxonomy, the conference is still attracting lots of newcomers to taxonomy and KM.
The Text Analytics Forum, now in its third year, also continues to expand. Of particular note are the case studies, which showcase a wide range of problems, approaches, and solutions.
Taxonomies & Ontologies & Knowledge Graphs
Yes, people are still talking about taxonomy, and there are still questions ranging from “What is taxonomy?” to “How do I deliver content in real-time using my taxonomy?”
What has changed, however, is the way taxonomists and non-taxonomists are talking about taxonomies.More and more, we are hearing about ontologies. Ontologies were once mainly in the purview of scientists and academics, but more and more, taxonomists need to have at least a working knowledge of RDF and some basic ontology principles. Ontologies are not simply more advanced taxonomies; they define knowledge domains and offer machine-readable, interoperable mechanisms for powering end-user applications like knowledge graphs. And here is the reason why people are talking more about ontologies: because they are all talking about knowledge graphs. In nearly all of the conferences, there were sessions addressing knowledge graphs and how they could power knowledge discovery and delivery.
Curiously, while knowledge management folks attending KMWorld 2019 have traditionally not talked much about taxonomies, they are very much talking about knowledge graphs. Regardless, they know that knowledge graphs are the next wave of delivering knowledge and taxonomies and ontologies are making those knowledge graphs possible.
Personalization & Knowledge Delivery
Tightly coupled with the rise of knowledge graphs is the potential they hold for personalizing knowledge discovery and delivery. While text analytics processes are identifying and extracting useful entities from structured and unstructured text, these entities are being semantically modeled in ontologies. In turn, these ontologies provide the logical model for knowledge graphs. Whether manually, automatically, or using a hybrid model, the content linked together by these entities and relationships can be used to present general information or can rely on machine learning and artificial intelligence to help curate a more personalized version of the content being delivered.
Knowledge management is all about personalization, knowledge discovery, and knowledge delivery. The use of knowledge graphs is allowing knowledge management efforts to achieve their goal of delivering the right knowledge to the right people at the right time. The more information the overall technology architecture has access to, the more it is possible to personalize and deliver content.
”Knowledge graphs are the next wave of delivering knowledge. Taxonomies and ontologies are making those knowledge graphs possible.
AI & Machine Learning
Some emerging and now dominant themes over the past few KMWorld conferences are around machine learning and artificial intelligence. Topics ranged from explanations about what these technologies are to whether or not they will replace human workers or particular tasks currently performed by human beings.
There remains a lot of uncertainty around machine learning and artificial intelligence due to their non-transparent nature and the amount of work it takes to train and perfect these systems. Thus, despite their frequency as session topics at KMWorld 2019, there are still no clear directions or applicable use cases which users can clearly see and understand.
Undoubtedly, though, these technologies are going to be inherent in the assembling of knowledge graphs and the delivery of knowledge.
Search will still be important to the discovery and delivery of knowledge, but the way the knowledge in the search results is assembled and presented will change. In general, there have not been any new revolutions in search, but the technologies aiding and supporting search have grown more robust. The delivery of information in a search interface is more likely to be tied together by knowledge graphs based on search concepts and natural language statements. These concepts and search terms match key concepts in labeled and semantically connected content, enriching the breadth of search results for end users.