ThinkBase

Few users of AI are aware of the complex history of the subject.

AI has two main sources of inspiration.

One is Nature; hence Neural Networks, Genetic Algorithms, Ant Trail Optimization and many other algorithms.

The other is language and logic: hence Expert Systems, Fuzzy Logic, classic NLP and others.

These two sources have fought it out over the years, with one dominant at one period and the other later on.

Right now, the nature thread is completely dominant. Deep learning is taught as the only form of AI, and all problems seem to have to have deep learning solutions.

The problem with deep learning is that the models generated are black boxes. This means that you cannot ask a deep learning model how it arrived at it's predictions, categorizations, decisions, etc.

This has given rise to the justified fear that our lives can be affected by the decisions of systems that may be making mistakes, but where their owners cannot explain, enumerate or remedy those mistakes.

It may not have escaped your attention that these deep learning models are all owned by the big tech companies. The bigger language models can cost tens of millions of dollars in compute time to train.

This explains the growth of interest in Knowledge Graphs as a technology that can both perform well and explain what it has learned and where it learned it from.

ThinkBase is a knowledge graph development and inference engine run as a SaaS.

It combines the ability to graph complex relationships as nodes, connections and attributes with the ability to make logical or heuristic inferences from those structures.

The logical and heuristic part is based on our Fuzzy Logic Inference Engine and knowledge definition language, DARL.

I first started working on the idea of combining graph theory and fuzzy logic in 1998. The story of how I fell out with the billionaire that was financing the development must be left to the history books. However, I always wanted to start again and this product is the result of that desire.

Andy Edmonds, PhD