A well-known axiom is that the quality of the machine learning algorithm is dependent on the quality of the examples from which it learns. We worked on creating high quality ground truth for a client in the legal business. There are multiple domains and ground truth creation in each of the domains was a manual process.
We accelerate the creation of the ground truth using our custom solution that uses best practices in Software Engineering, UI and Data Science. We manage tagging by multiple users through a friendly user interface and do an automated management of conflicting tags. We also deployed learning algorithms to take the tagged documents and automatically assign categories to the untagged documents.
70% – Reduction in Tagging time
99% – Accuracy in automatically tagged documents
1M+ – Cost Savings