New GREP framework improves semiconductor manufacturing process monitoring

Researchers have developed a new self-supervised learning framework called Global Regression Pretraining (GREP) for predicting the material removal rate (MRR) in chemical mechanical planarization (CMP), a crucial semiconductor manufacturing process. GREP utilizes attention mechanisms and learns temporal representations from unlabeled sensor data through a global proxy-target regression task before being fine-tuned for MRR estimation. The framework demonstrates competitive performance, particularly under conditions with limited labeled data, and shows improvements in handling sensor noise and temporal process drift compared to a feature-engineered supervised baseline. AI

IMPACT This self-supervised learning approach could enhance efficiency and reduce reliance on physical metrology in semiconductor manufacturing.

RANK_REASON The cluster contains an academic paper detailing a new methodology for a specific technical problem. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• chemical mechanical planarization
• Global Regression Pretraining
• Hugging Face
• Jonathan Adam Rico
• material removal rate
• PHM 2016 CMP dataset
• semiconductor manufacturing

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →