filmov
tv
Mathematical Optimization + Artificial Intelligence = MOAI (English AI Talk Version)
![preview_player](https://i.ytimg.com/vi/ms9PqhJr1SQ/maxresdefault.jpg)
Показать описание
I would like to talk to you about "Mathematical Optimization + Artificial Intelligence = MOAI."
MOAI is a term I coined to represent the fusion of Mathematical Optimization and Artificial Intelligence, as the name suggests.
In this presentation, I first proposed seven classifications for the field of Machine Learning and Mathematical Optimization Integration (MOAI). These classifications are not exhaustive but serve as a convenient guideline. MOAI has been studied independently across various fields, and moving forward, cross-disciplinary collaboration will be essential. I also proposed a new framework (Encode-Decode method) that uses past problem instances and solutions as teachers. Additionally, I introduced a new framework (MO Hybrid) for dynamic optimization models. I highlighted the need for fundamental research on the distance and similarity between problem instances and provided some basic ideas. Experimentation and application of these methods to real-world problems remain as future challenges. Lastly, I would appreciate it if practitioners who repeatedly solve optimization problems could store many instances and solutions.
#optimization #AI ##数理最適化 #機械学習 #MOAI
@サプライ・チェイン最適化チャンネル(MIKIO KUBO)
動画プレイリスト:
Juliaで数理最適化
MOAI is a term I coined to represent the fusion of Mathematical Optimization and Artificial Intelligence, as the name suggests.
In this presentation, I first proposed seven classifications for the field of Machine Learning and Mathematical Optimization Integration (MOAI). These classifications are not exhaustive but serve as a convenient guideline. MOAI has been studied independently across various fields, and moving forward, cross-disciplinary collaboration will be essential. I also proposed a new framework (Encode-Decode method) that uses past problem instances and solutions as teachers. Additionally, I introduced a new framework (MO Hybrid) for dynamic optimization models. I highlighted the need for fundamental research on the distance and similarity between problem instances and provided some basic ideas. Experimentation and application of these methods to real-world problems remain as future challenges. Lastly, I would appreciate it if practitioners who repeatedly solve optimization problems could store many instances and solutions.
#optimization #AI ##数理最適化 #機械学習 #MOAI
@サプライ・チェイン最適化チャンネル(MIKIO KUBO)
動画プレイリスト:
Juliaで数理最適化