POLISHING DEVICE, INFORMATION PROCESSING SYSTEM AND PROGRAM

To enhance throughput or to improve yield by preventing a defective from being missed.SOLUTION: A polishing device 10 is capable of referring to a storage in which machine learning models are stored, while inputting a feature amount of a signal relating to a friction force between a polishing member (a polishing pad 101) under polishing and a wafer W or a feature amount of a temperature of the polishing member under polishing or of the wafer and using data relating to a thickness of the wafer after polishing or parameters relating to yield of a product included in the wafer after polishing. The polishing device comprises: a polishing table 100 in which the polishing member is provided; a control unit which performs control in such a manner that the wafer is pressed to the polishing member and polished, while rotating the polishing table and a polishing head 1 opposed to the polishing table; and an AI unit which generates the feature amount from the friction force between the polishing member and the wafer or from the temperature of the wafer and inputs the feature amount to a learnt machine learning model, thereby outputting as an estimate the data relating to the thickness of the wafer after polishing or the parameters relating to the yield of the product included in the wafer after polishing.SELECTED DRAWING: Figure 1 【課題】不良品を見逃さずにスループットを改善するまたは歩留まりを向上する。【解決手段】研磨装置１０は、研磨中の研磨部材（研磨パッド１０１）と基板Ｗとの間の摩擦力に関する信号の特徴量又は研磨中の研磨部材もしくは基板の温度の特徴量を入力とし、研磨後の基板の膜厚に関するデータ若しくは研磨後の基板に含まれる製品の歩留まりに関するパラメータを用いて機械学習モデルが記憶されているストレージを参照可能であり、研磨部材が設けられ研磨テーブル１００と、研磨テーブルに対向する研磨ヘッド１とを回転させながら、基板を研磨部材に押し付けて研磨するよう制御する制御部と、研磨部材と基板の間の摩擦力又は基板の温度から特徴量を生成し、特徴量を学習済みの機械学習モデルに入力することで、研磨後の基板の膜厚に関するデータ又は研磨後の基板に含まれる製品の歩留まりに関するパラメータのいずれかを推定値として出力するＡＩ部と、を有する。【選択図】図１