Outlier detection with ML precision.
Many chipmakers face a difficult trade-off — improve quality without affecting yield. Traditional testing methods fail to navigate this challenge due to their limited visibility below the pass/fail limits, discarding perfectly good chips or letting small defects slip through to the field. The challenge is clear: manufacturers must achieve both quality and yield goals without sacrificing one for the other.
This paper introduces proteanTecs groundbreaking Outlier Detection solution that eliminates that tradeoff. proteanTecs’ Outlier Detection uses deep data analytics and ML to detect latent defects as early as Wafer Sort, achieving high fault detection accuracy by learning normal behavior with on-chip agents and comparing test measurements with predicted ones. It identifies marginal issues beyond simple pass/fail metrics, where traditional methods fail. To read more, click here.
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
Leave a Reply