[3F5-GS-10] AI application: Manufacturing

[3F5-GS-10-01] A Note on Similar Case Retrieval via Deep Metric Learning Using Sensor Data Obtained from Semiconductor Manufacturing Equipment

〇Naoki Saito¹, Ren Togo¹, Keisuke Maeda¹, Ruiki Kobayashi², Takahiro Nakamura², Motohiro Okaya², Masato Kazui², Takahito Matsuzawa², Takahiro Ogawa¹, Miki Haseyama¹ (1. Hokkaido University, 2. Tokyo Electron Limited)

Keywords:Deep metric learning, Semiconductor manufacturing equipment, Sensor data, Similar case retrieval

This paper presents a similar case retrieval method based on deep metric learning using sensor data from semiconductor manufacturing equipment. In semiconductor manufacturing, early detection of issues and swift recovery are crucial for improving the operation rate of the equipment. To achieve this, it is expected to realize an equipment condition monitoring technique using sensor data and a similar case retrieval technique to identify the anomaly type that has occurred and propose a remedial action. The proposed method realizes similar case retrieval using feature representations obtained from sensor data of semiconductor manufacturing equipment using a deep metric learning model. Experimental results using sensor data obtained from semiconductor manufacturing equipment show that the proposed method can retrieve similar cases with the same anomaly type as the query case with high precision.