A machine learning approach to yield management in semiconductor manufacturing

Yield improvement is one of the most important topics in semiconductor manu facturing. Traditional statistical methods are no longer feasible nor effic ient, if possible, in analysing the vast amounts of data in a modern semico nductor manufacturing process. For instance, a typical wafer fabrication pr ocess has more than 1000 process parameters to record on a single wafer and one manufacturing plant may produce tens of thousands wafers a day. Tradit ional approaches have limits in extracting the full benefits of the data. T herefore, the manufacturing data is poorly exploited even in the most sophi sticated processes. Now it is widely accepted that machine learning techniq ues can provide powerful tools for continuous quality improvement in a larg e and complex process such as semiconductor manufacturing. In this work, me mory based reasoning (MBR) and neural network (NN) learning are combined fo r yield improvement and an integrated framework is proposed for a yield man agement system based on hybrid machine learning techniques. In this hybrid system of NN and MBR, the feature weight set which is calculated from the t rained neural network plays the core role in connecting both learning strat egies and the explanation on prediction can be given by obtaining and prese nting the most similar examples from the case base. The proposed system has advantages in typical semiconductor manufacturing problems such as scalabi lity to large datasets, high dimensions and adaptability to dynamic situati ons.