P2P

92 P E E R T O P E E R : I L T A ' S Q U A R T E R L Y M A G A Z I N E | W I N T E R 2 0 2 3 precision of AI-driven analysis in legal document review through all the upgrades made to KIBIT in 2022. The graph indicates that to reach a level of 80% recall (which is equivalent to finding 80% of the relevant documents) with the improved algorithms, a human reviewer would be required to go through only approximately 3% of the entire data set vs about 18% with the older algorithms. Algorithmic Innovation FRONTEO's priorities with AI development remain focused on efficiency and accuracy of its "KIBIT" engine, and FRONTEO's sizable development staff and data science experts are working on bringing a new algorithm to market in 2023. The original algorithm behind "KIBIT" in 2012 relied on feature weighting to identify relevance in a data set in an effective manner with a small amount of training data and few positive training examples. The team innovated in 2019, developing a neural-network random forest-based algorithm to deliver extreme precision in datasets with medium to high richness, which had been a great success in delivering results to clients in a variety of cases. FRONTEO is now bringing to market a new algorithm that combines the accuracy and low training data requirement of our previous algorithms. This new algorithm was inspired by intuitive cognition found in humans and simulates the human ability to make snap judgments about color composition based on the hues of its individual components in a picture (see Figure 2) This algorithm will enable AI to score and rank the text of a document based on the rankings of individual morphemes found within those documents quickly and effectively. It also incorporates the various know-how and technological breakthroughs obtained in the development of the previous two algorithms used by KIBIT, combining the best of all worlds, and enhancing the features of each. F E A T U R E S This new algorithm will be a general-purpose, high-performance AI engine that can serve as a singular go-to algorithm for providing fast and accurate results in CJK languages. It will also feature a built-in display for confidence interval around the recall curve to improve defensibility of the AI review process. F I G U R E 2 : E A C H W O R D I S C O L O R E D A C C O R D I N G T O R E L E V A N C E , C R E AT I N G A N O V E R A L L F E E L F O R T H E D O C U M E N T. F I G U R E 3 : C O N F I D E N C E I N T E R V A L A R O U N D A R E C A L L E S T I M AT E I N T H E N E W A L G O R I T H M

• Cover