Defect Prediction in Lithography using MLP

Sample Use Case

Classifier system to predict whether or not a certain patterning defect mode will occur based on two inline measurements

  • Scanner exposure dose
  • Focus.

This defect mode is typically only detected in post-etch inspection. However, if we could predict whether the defect will occur immediately after the exposure step (based on available scanner data) we could simplify the rework process and avoid potential yield loss.

For first pass we have good metrics to with Precision, Recall and F1 as 98 %

Use Case – Edge Analytics Hotspot Identification

A mismatch between lithography wavelength and semiconductor technology feature size causes diffraction when transferring the layout from design onto a silicon wafer.

Resulting in  open or short circuits i.e. lithography hotspots.

We leverage Vision Transformers to identify Hotspots, with and accuracy of 98%

Sample Use Case: Supply Chain Optimization

Contextual optimization of supply chain using Linear Programming , considering

  • Freight Cost
  • Fixed and Variable Costs
  • Socio Economic Impact – e.g. Covid /CO2 Emissions
  • Delivery Time , Lead Time
  • Capacity
  • Solution takes into consideration dynamic variables and performs linear optimization towards objective function – for how to calibrate production. 

Use Case – Total Security

Security Incident and Event Monitoring –

  • Threat Intelligence
  • Security Suite
  • Network
  • Alert
  • Application Server
  • System Events
  • Event Taxonomy
  • Guest Access
  • Database Control
  • Identity Management
  • File Access Control

Threat Intelligence Dashboard

Classification of threat messages, external malicious IP addresses, suspicious malicious activity counts.

Why edge

  • Bandwidth and Latency
  • Security and Decentralization
  • Swarm intelligence
  • Cost effective

Ever evolving

1.Economies of scale FPGAs/ASICs/GPUs/TPUs

2.Reduced algorithmic complexity – Approximate Computing – Quantized Neural Networks , Hardware Co-optimization – In memory computing