top of page
sub02_visual_img.jpg
Cyber Physical System

Singe-S AI-VM

Real time self learning-based process capability

prediction server Reinforcement learning engine of

Signe-S VM makes accurate predictions even when processes change

Virtual Metrology
Singe-S AI-VM

Main Features

  • Easy initial setup to reduce deployment costs

  • AI autonomously monitors process changes

  • AI automatically extracts the cause of process change in real time

  • AI autonomously evolves models with self-learning

  • AI autonomously manages the accuracy of virtual metrology

Benefit

  • Quality can be prevented by accurately predicting measurement results in advance

  • It allows to achieve productivity improvement and quality at the same time by measuring only in situations where measurement is absolutely necessary.
    (Smart Measurement Sampling)

  • Reduce the purchase of expensive metrology

  • Quality analysis and process control capabilities are improved due to sufficient measurement data

Core functions
sub02_img_05_e.png
Autonomous detection
of process changes
Autonomous
model evolution
Self-learning after real-time autonomous
detection of multivariate changes
I extracts the timing of process change and the root cause that can cause a decrease in virtual metrology accuracy

Artificial intelligence detects autonomously in real-time

Advanced algorithm for multivariate change
detection

Automated extraction of root cause in real-time

(Variable Contribution)

sub02_img_02_update.png
Real-time
predictions
Autonomous
management of
predictive accuracy
Self-management to monitor anomalies
in prediction "Explainable AI"
AI extracts the timing of process change and the root cause that can cause a decrease in virtual metrology accuracy

Model Simulation

Predicted SPC
Relevance Monitoring

Prediction Reliability
Monitoring & Alarm

Explain the rationale for autonomous
model evolution

multivariate change analysis, variable analysis, residual
analysis, performance analysis, etc.

Prediction Performance
Monitoring

Signe-S VM enables accurate prediction even under process changes that are not possible with existing VMs, enabling hyper automation

sub02_img_06.png

From semiconductors, batteries, chemicals, electricity, electronics, machinery, distribution, to food and beverages, precise real-time predictions enable predictive management across these industries.

bottom of page