Knowledge graphs are like interactive and automated feature charts of patents. They make possible for the computer to reach and exceed human level understanding of technology.
All technical features and their essential relationships and interactions are expressed in a natural and visual way.
And technology, like mechanics, is all about relationships and interactions.
In electronics, the terminology used is very similar from patent to another. The functional connections between elements are crucial.
The connections can be powerfully expressed using graphs to enable accurate patent searching and classification.
Software related patents are hard to search for. The same terms keep repeating in millions of patents, the differences being in nuanced functional details in ponderous texts.
Graph AI is a great companion with software. It can build understanding of complex technical operations and find matching patents - without getting exhausted.
In chemistry, there are a lot of chemical names, attributes and numerical values and ranges.
Our graph parser extracts all those to an easy-to-read format and the AI is trained with real-life claims and cited prior art to interpret and understand their meaning.
