Objective of the integrated circuit testing is to identify the defective chips before shipping to the customers  and ensure that a certain test quality level is met at an economically acceptable cost. However,  ever increasing complexity of integrated circuits with each new manufacturing process generation necessitates costlier test sets to meet test quality goals and it becomes increasingly difficult to find a balance between them. 

An identical test set is traditionally applied to all manufactured chips during the test process. This project develops a data-driven adaptive test methodology that dynamically adjusts the test set during the test process based on the changing defect population with a goal of test optimization. The proposed approach tackles this problem by modeling  instantaneous  test quality and test cost relation by real-time analysis of test data. The model is utilized to explore test quality and test cost tradeoff space and subsequently optimal cost test set that delivers the target test quality is identified. 

A set of methodologies that enables the systematic evaluation of possible test quality-test cost models and the real-time adaptation of models based on evolving defect characteristics are developed throughout the project. Finally, an alternative use of test quality-test cost model to optimize the test cost across different test steps and the use of proposed methods in practice are explored.