The typical introduction steps will be as follows.   

1.  Conclude an NDA.
   

2. We will disclose the technical information and you will provide us with a data set for initial evaluation measured with your existing Source Measurement Units (SMUs).  Through candid discussions, we will develop a mutual understanding of the measurement environment, measurement processes, your strategy and concerns, etc. 

   • The number of data at this time will be 20 to 30 sets.
     

   •  The data will be  2 minutes long according to our recipe.
     

   • We analyze the sample data set and will feedback.
     
     

3.   There is no case without problems with data resolution, noise, missing data, or errors.
   

    

   

   •  If a solution is required, we will discuss it with you.
     

   •  If the problem cannot be solved in-house, the following engineering services are also available.
     

   • In many cases, existing SMUs can be used.
     

4.  Once we could confirm that the data sets are available, you will start to measure 1000 DUTs.
   

5. Half of the sample dataset is used to actually build thealgorithm, while the other half is used to validate the prediction accuracy.
   

   • Training data and Test or Validation data are not duplicated but are used in a random 50/50 split.
     
     

6.  We create algorithms for you.  

   •  The algorithms are unique and optimized for your measurement environment.
     

   •  You evaluate whether it meets the target prediction accuracy.
     

7. Conclude a master service agreement and ancillary agreements.