AI Averbis - "not relevant" is taken into account more specifically in training

More precise AI results through controlled training

To support weekly monitoring, Averbis can be used to carry out training based on relevant and "non-relevant" patents can be carried out. The aim is to identify "irrelevant" documents at an early stage and sort them out automatically.

The system continuously improves its detection quality through machine learning.

With the current adjustment, the Targeted increase in the quality of training data:

• Only "non-relevant" patents Are taken Into account that have been manually classified as such by users.
• Automatically set or uncertain Assignments are no longer included in The training.