Using AI For Fault Detection And Classification In Semiconductor Manufacturing

Using AI For Fault Detection And Classification In Semiconductor Manufacturing

by Semiconductor Engineering

Applied Sciences & Engineering Engineering & Technology TGPProduction and industrial engineeringTGPQIndustrial quality control
🎙 Semiconductor Engineering 👥 30K 📅 September 5, 2025 ⏱ 12 min 👁 2K 📄 expert opinion 🧭 2026-08-17
Available in: English (current) Français

Keywords

FDCmachine learningsupervised learningunsupervised learninganomaly detection

Summary

In this interview, Ed Berling of Semiconductor Engineering speaks with Jon Herlocker, CEO of Tignis (now part of Cohu), about the application of AI to fault detection and classification (FDC) in semiconductor manufacturing. Herlocker contrasts classic FDC, which relies on manual threshold setting and is reactive, with machine learning-based approaches that can process vast sensor data and detect complex patterns. He emphasizes the limitations of supervised learning, which requires historical fault examples, and highlights the value of unsupervised learning for anomaly detection, especially for novel faults. The discussion covers the challenges of data preparation, model deployment, and the importance of explainability in root cause analysis. Herlocker also mentions automated correlational analysis as a tool to identify relevant sensors and understand yield variations. The video is part of a seven-part series on AI in manufacturing, with the next episode on virtual metrology.

141 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides valuable insights into the practical application of machine learning for FDC, highlighting the shift from reactive to predictive maintenance. The argumentation is solid, grounded in the speaker’s industry experience, and clearly explains the differences between supervised and unsupervised learning in this context. The discussion on data preparation and model deployment addresses real-world challenges, making the content highly relevant for professionals in semiconductor manufacturing.

Scientific Rigor, Source Quality, Title Accuracy

The video is an expert interview without formal citations, but the speaker’s position as CEO of Tignis lends credibility. The content aligns with known industry trends and practices. The title accurately reflects the content, and the discussion is technically rigorous, though it could benefit from more concrete examples or data. No comments were provided for analysis.

137 words

Title / Content Match

The title accurately reflects the content, which focuses on using AI for fault detection and classification in semiconductor manufacturing.

Quality & Reliability

8/10

The video features an expert interview with the CEO of Tignis, providing credible insights into the application of machine learning for FDC in semiconductor manufacturing. The discussion is technically sound and aligns with industry practices, though it lacks detailed data or citations.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Contribution & Novelties

The video offers a clear explanation of how unsupervised learning can address the limitations of classic FDC, particularly for detecting novel faults. It also highlights the importance of data preparation and model deployment, which are often overlooked. The discussion on automated correlational analysis provides a practical approach to sensor selection.

Pour aller plus loin :

  • Anomaly detection — Overview of anomaly detection techniques.
  • Machine learning in semiconductor manufacturing — Related article on AI applications.
  • Explainable AI — Concept of explainability in AI.

82 words

Radar Profile

The radar profile shows high scores in quality and reliability, with moderate scores in quantity and technical level, indicating a focused and credible discussion suitable for professionals.

Reliability 8/10