Thanks for them Knowledge Graphs

It's Thanksgiving Day here in the US. A time to count one's blessings.

I'm thankful for your attention and discussions and disagreements.

I passed two milestones not long ago: more than 5,000 (now almost 6,000!) followers and more than 3,000 subscribers -- most of them in the last 6 months. Discussions around my posts have led to many new learnings and many fascinating new contacts -- and to a wave of new business opportunities. This, in turn, is pushing me to build a very talented team (more on that in the New Year).

I'm thankful that knowledge graphs are gaining so much momentum.

Gartner already rates knowledge graphs as critical enablers in 2024. They're here. Now.

They're needed immediately. In data integration layers, in graph RAG solutions, for training next-gen LLMs, systematizing data aggregation, as infrastructure for reliable features that drive better machine learning -- and as an essential cornerstone for Safe AI as well as for mitigating a wide range of other business risks. Knowledge graphs are essential for taming the vast jungle of unstructured text data (I'm giving a full-semester course on Big Text Data starting in January) and to power more expert-like reasoning (a visiting lecture on this at Oxford is coming up) -- all so we can offer much better search results and recommendations, for example.

Sure, not all of the knowledge graphs are as rich and densely interconnected as we would like. But there's been an amazing amount of wonderful work this year on automating how we build them, on ways of enriching the impoverished ones, and on developing applications to show their dramatic impact. Our KG community is figuring out how to build them from text, from databases, from terminologies, from taxonomies -- which will accelerate progress.

And I'm thankful we live in interesting times.

We have powerful new tools in our hands today. We can and will turn them into powerful new opportunities. Looking forward to next adventures.