Knowledge graphs bring patent searches to the new levels. Instead of reading text, our deep learning models ingest compact graphs that contain not only technical features but also their essential relationships in a natural way. This brings a whole new dimension to patent searching.
The neural network behind the search is trained with millions of real-world patent cases and has learned to imitate the work of patent examiners. Except for the speed! In only a couple of seconds you will get the technically most relevant hits out of tens of millions of patent publications.
Find out quickly why the search suggests that one exotic result! The graph approach allows you to see what the machine is thinking and makes it easy and fun to investigate the results. You save time and get a better understanding of the relevant content of prior art.
Besides building the most powerful search engine ever, we are developing other game-changing patent evaluation and preparation tools. Being inspired by both complex problems and opportunities beyond them, we invite you to join us as a problem-solver or collector of the benefits.