DEEP REINFORCEMENT LEARNING-BASED INTEGRATED CIRCUIT DESIGN SYSTEM AND METHOD USING PARTITIONING

The present invention may provide parameterized hyperparameter partitioning in consideration of the balance of a partition size while preserving a hypergraph property necessary for applying deep reinforcement learning by reducing a large size hypergraph, and can reduce the amount of calculation and...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: LE, Pham-Tuyen, YOON, Do-Kyoon
Format: Patent
Sprache:eng ; fre ; kor
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The present invention may provide parameterized hyperparameter partitioning in consideration of the balance of a partition size while preserving a hypergraph property necessary for applying deep reinforcement learning by reducing a large size hypergraph, and can reduce the amount of calculation and capacity of an artificial neural network through graph reduction. La présente invention peut fournir un partitionnement d'hyperparamètre paramétré en tenant compte de l'équilibre d'une taille de partition tout en préservant une propriété d'hypergraphe nécessaire pour appliquer un apprentissage de renforcement profond par réduction d'un hypergraphe de grande taille, et peut réduire la quantité de calcul et de capacité d'un réseau neuronal artificiel par réduction de graphe. 본 발명은 큰 크기의 하이퍼 그래프를 축소하여 심층 강화학습을 적용하기에 필요한 하이퍼 그래프의 성질을 보존하면서 파티션 크기의 균형을 고려한 매개화된 하이퍼 파라미터 파티셔닝을 제공할 수 있으며 그래프 축소를 통해 인공 신경망의 연산량과 용량을 축소시킬 수 있다.