岡山理科大学マネジメント学会 第9回研究会
岡山理科大学マネジメント学会では、岡山統計研究会との共催で、シンガポール及びドイツから来日されるデータサイエンスの分野の研究者の方をお迎えし、深層学習や計算環境をテーマとして、第9回研究会をハイブリッド方式で開催します。
ご興味ある方、ぜひご参加くださいますようご案内申し上げます。
第9回研究会 開催案内
日 時:2022年9月22日(木)16:25~
場 所:岡山理科大学 A1号館1階 プレゼンテーションルーム
開催方法:Zoom併用のハイブリッド方式(参加費無料)
- Zoom IDは、下記申込でオンラインを選択された方にお知らせします。
- 対面参加をされる方は、感染拡大防止について、ご協力をお願いします。会場では、消毒液の配置、換気等に留意しますが、個人におかれましても、マスクの着用や手指消毒、体調管理、感染リスクのある行動の回避等をよろしくお願いします。
テーマ:Deep learning and Computational aspects
申し込み:こちらからお申し込みください(Google Form)。
なお、申込なしでも当日ご参加いただけますが、オンラインの場合は必ずお申し込みください。
共 催:岡山統計研究会
問合せ先:岡山理科大学経営学部 森 裕一<yuichi-moriあっとous.ac.jp>
プログラム:
16:25 開会
16:30 講演1
17:30 講演2
18:30 総合討論
Lecture 1:
Deep Switching State Space Model (DS3M) for Nonlinear Time Series Forecasting with Regime Switching
Xiuqin Xu, *Ying Chen (National University of Singapore, Singapore)
Abstract: We propose a deep switching state space model (DS3M) for efficient inference and forecasting of nonlinear time series with irregularly switching among various regimes. The switching among regimes is captured by both discrete and continuous latent variables with recurrent neural networks. The model is estimated with variational inference using a reparameterization trick. We test the approach on a variety of simulated and real datasets. In all cases, DS3M achieves competitive performance compared to several state-of-the-art methods (e.g. GRU, SRNN, DSARF, SNLDS), with superior forecasting accuracy, convincing interpretability of the discrete latent variables, and powerful representation of the continuous latent variables for different kinds of time series. Specifically, the MAPE values increase by 0.09% to 15.71% against the second-best performing alternative models.
Lecture 2:
Progress in Mathematical Programming Solvers from 2001 to 2020
*Thorsten Koch (Zuse Institute Berlin & Technische Universitaet Berlin, Germany)
Abstract: We report on a study that investigates the progress made in LP and MILP solver performance during the last two decades by comparing the solver software from the beginning of the millennium with the codes available today. On average, we found out that for solving LP/MILP, the total speed-up was about 180 and 1,000 times, respectively. However, these numbers have a very high variance and they considerably underestimate the progress made on the algorithmic side: many problem instances can nowadays be solved within seconds, which the old codes are not able to solve within any reasonable time. We will report on how we measure performance and why it is very difficult to come up with one reasonable number.
Discussion on Analysis for Big Data and Complex data
Discussants:
Ying Chen, National University of Singapore, Singapore
Thorsten Koch, Zuse Institute Berlin & Technische Universitaet Berlin, Germany
Ralf Borndörfer, Zuse Institute Berlin and Freie Universität Berlin, Germany
Yuji Shinano, Zuse Institute Berlin, Germany
Uwe Gotzes, Open Grid Europe (OGE) , Germany
Masaya Iizuka, Okayama University, Japan
Masahiro Kuroda, Okayama University of Science, Japan
Yuichi Mori, Okayama University of Science, Japan