Classification and Localization of Semiconductor Defect Classes in Aggressive Pitches (imec, Screen)

A new technical paper titled “An Evaluation of Continual Learning for Advanced Node Semiconductor Defect Inspection” was published by Imec and SCREEN SPE Germany.

Abstract

“Deep learning-based semiconductor defect inspection has gained traction in recent years, offering a powerful and versatile approach that provides high accuracy, adaptability, and efficiency in detecting and classifying nano-scale defects. However, semiconductor manufacturing processes are continually evolving, leading to the emergence of new types of defects over time. This presents a significant challenge for conventional supervised defect detectors, as they may suffer from catastrophic forgetting when trained on new defect datasets, potentially compromising performance on previously learned tasks. An alternative approach involves the constant storage of previously trained datasets alongside pre-trained model versions, which can be utilized for (re-)training from scratch or fine-tuning whenever encountering a new defect dataset. However, adhering to such a storage template is impractical in terms of size, particularly when considering High-Volume Manufacturing (HVM). Additionally, semiconductor defect datasets, especially those encompassing stochastic defects, are often limited and expensive to obtain, thus lacking sufficient representation of the entire universal set of defectivity. This work introduces a task-agnostic, meta-learning approach aimed at addressing this challenge, which enables the incremental addition of new defect classes and scales to create a more robust and generalized model for semiconductor defect inspection. We have benchmarked our approach using real resist-wafer SEM (Scanning Electron Microscopy) datasets for two process steps, ADI and AEI, demonstrating its superior performance compared to conventional supervised training methods.”

Find the technical paper here.  Published July 2024 (preprint).

Prasad, Amit, Bappaditya Dey, Victor Blanco, and Sandip Halder. “An Evaluation of Continual Learning for Advanced Node Semiconductor Defect Inspection.” arXiv preprint arXiv:2407.12724 (2024).