The Next Generation Document Review Tool
SENTIO’s Analytic Review CAL tool provides:
- Simultaneous training of multiple models (tag multiple issues)
- No need for seed sets
- Continuous reshuffling of documents
- Real‐time monitoring of model building
- Saving of models for use in subsequent productions or matters/cases
Continuous Active Learning/CAL
SENTIO Software uses machine learning to find relevant documents based on input provided by human users. Analogous to a music streaming service “choosing” which songs a user will enjoy based on previous song selections, SENTIO Software uses document tags to train a ranking algorithm that orders the documents from most to least likely to be relevant.
Unlike traditional linear review, SENTIO Software allows the managing attorney to determine that a review is complete without requiring “eyes‐on” review of every document. By a user reviewing and tagging a small portion of the data set the software creates a training model which sorts “like” documents to the top of the population to automatically generate predictive results.
SENTIO Software provides clients real‐time reporting that shows the accuracy of the trained model on‐the‐fly. This allows users to decide immediately if the training process should continue or not. Once the effective model has been trained, the tagging can be propagated through the remaining document collection. In this way, SENTIO Software provides lawyers with the ability to algorithmically eliminate human review of a large percentage of a document collection—often by as much as 90%, during the first-pass review, potentially reducing a company’s document review and management expense by millions of dollars and achieving improved results in far less time.