OPTIMAL SOLUTION ACQUISITION PROGRAM, OPTIMAL SOLUTION ACQUISITION METHOD, AND INFORMATION PROCESSING DEVICE

To reduce the time required for acquiring an optimal solution.SOLUTION: An information processing device performs learning of a variational auto-encoder using multiple training data including an objective function. The information processing device inputs multiple training data to a trained variatio...

Full description

Saved in:
Bibliographic Details
Main Author OTA EIJI
Format Patent
LanguageEnglish
Japanese
Published 11.01.2022
Subjects
CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:To reduce the time required for acquiring an optimal solution.SOLUTION: An information processing device performs learning of a variational auto-encoder using multiple training data including an objective function. The information processing device inputs multiple training data to a trained variational auto-encoder to identify a distribution of the multiple training data in a latent space of the trained variational auto-encoder. The information processing device determines a search range for an optimal solution of the objective function according to the distribution of the multiple training data, and acquires the optimal solution of a desired objective function by using the training data included in the determined search range.SELECTED DRAWING: Figure 15 【課題】最適解の獲得にかかる時間を短縮することを課題とする。【解決手段】情報処理装置は、目的関数を含む複数の訓練データを用いて、変分オートエンコーダの学習を実行する。情報処理装置は、学習済みの変分オートエンコーダに複数の訓練データを入力して、学習済みの変分オートエンコーダの潜在空間上における複数の訓練データの分布を特定する。情報処理装置は、複数の訓練データの分布に従って、目的関数の最適解の探索範囲を決定し、決定された探索範囲に含まれる訓練データを用いて、所望する目的関数の最適解を獲得する。【選択図】図１５
Bibliography:Application Number: JP20200107713