第1回 RWBC-OIL Workshop
産総研・東工大 実社会ビッグデータ活用 オープンイノベーションラボラトリ(RWBC-OIL)では、 平成29年2月20日の設立以来、実社会ビッグデータ活用技術による新たな価値創造を実現するために、 産総研と東工大が有する計算プラットフォーム構築技術とビッグデータ処理技術を融合し、さまざまな 分野に適用できるビッグデータの処理・解析技術を提供するオープンプラットフォームを構築することで、 新たな価値を創造するための研究開発を行ってきました。
RWBC-OILの活動ならびに研究成果を紹介するRWBC-OIL Workshopを開催いたします。 RWBC-OILでは、民間企業と密接に連携し共同研究や技術移転を進めることで、得られた成果の速やかな 産業化と社会実装を目指しています。 皆様のご参加をお待ちしております。
日 時:2018年5月8日(火)10:30〜17:45 (10:00受付開始)
場 所:
東工大蔵前会館(東京都目黒区大岡山2丁目12-1)
参加費:無料 主 催:産総研・東工大 実社会ビッグデータ活用 オープンイノベーションラボラトリ
参加申込み:
こちらから
RWBC-OIL Workshop Program
Date: Tuesday, May 8th, 2018
Venue: Kuramae Hall, Tokyo Tech Front (2-12-1 Ookayama Meguro, Tokyo)
Chair: Dr. Hirotaka Ogawa, Prof. Toshio Endo
Time Table
10:30 - 10:40 Opening remarks
Dr. Satoshi Sekiguchi, Prof. Osamu Watanabe
10:40 - 10:45 RWBC-OIL overview
Dr. Kazuhito Yokoi
10:45 - 11:00 RWBC-OIL's expectations
Prof. Katsuki Fujisawa
11:00 - 11:15 ABCI: AI Bridging Infrastructure
Dr. Hirotaka Ogawa
11:15 - 11:20 Q&A
11:20 - 11:40 Invited talk 1: Zebrafish Neural Activity Maps for Novel Neuromorphic Deep Learning Architectures
Dr. Gerald Pao, University of California, San Diego, USA
11:40 - 11:50 Q&A
11:50 - 12:40 Lunch
12:40 - 13:20 Invited talk 2: PPP-Net: Platform-aware Progressive search for Pareto-optimal Neural Architectures
Prof. Min Sun, National Tsing Hua University, Taiwan
13:20 - 13:30 Q&A
13:30 - 13:40 Overview of RG1
Dr. Ryousei Takano
13:40 - 13:55 Characterizing the Interference between I/O and MPI Traffic on Fat-tree Networks
Kevin Brown
13:55 - 14:10 Speeding Up Deep Learning with Second Order Optimisation and Distributed Training
Yohei Tsuji
14:10 - 14:25 RWBC-OIL Activities for Data-Driven Optimization of Datacenter Operation
Dr. Shinichiro Takizawa
14:25 - 14:30 Q&A
14:30 - 14:40 Overview of RG2-1
Prof. Koichi Shinoda
14:40 - 14:55 Deep learning for Remotely Sensed Data Classification/Segmentation
Dr. Poliyapram Vinayraj
14:55 - 15:10 Compressing Deep Neural Networks
Choong Jun Jin
15:10 - 15:15 Q&A
15:15 - 16:15 Poster session
16:15 - 16:30 Overview of RG2-2
Dr. Takao Terano
16:30 - 16:35 Q&A
16:35 - 16:45 Overview of RG2-3: Ultra-High Dimensional Data Analysis
Dr. Jun Sese
16:45 - 17:00 MEGADOCK: a supercomputing bioinformatics application for protein-protein interaction prediction
Dr. Masahito Ohue
17:00 - 17:15 Classifying Phenotype from Genotype via Deep Learning Models
Dr. Tony Kuo
17:15 - 17:20 Q&A
17:20 - 17:35 Overview of RG2-4: Big Data and Language Grounding
Dr. Hiroya Takamura
17:35 - 17:40 Q&A
17:40 - 17:45 Closing
Poster session
Mohamed Wahib, protoNN: Framework Agnostic, Scalable, and Optimization-driven Prototyping of DNNs for GPU-accelerated Supercomputers
Chen Peng, Pushing the Limits for 2D Convolution Computation On CUDA-enabled GPUs
Mateusz Bysiek, Between Python and Performance
Kazuaki Matsumura, Autotuning Temporal Blocking via Polyhedral Compilation
Tianqi Xu, HuronFS: Hierarchical, User-level and On-demand Burst Buffer File System
Yohei Tsuji, Fault Oblivious Distributed Deep Learning Applications
Shinichiro Takizawa, Preliminary Data Analysis for Efficient Deep Learning Job Scheduling
Shweta Salaria, Predicting Performance Using Collaborative Filtering
Kevin Brown, Co-locating Graph Analytics and HPC Applications
Ryo Matsumiya, vGASNet: Scalable RMA-based Communication Library for Out-of-core Data Processing
Jian Guo, Predictions for Underestimation of Job Runtime on HPC System using Machine Learning
Poliyapram Vinayraj, Deep learning for medium resolution satellite image classification
Yuya Murata, Development of Analysis Method for Corporate Data
Kento Aoyama, Comparison of the HPC Container Technologies for Bioinformatics Application
Tomohiro Ban, Efficient Hyperparameter Optimization by Using GP-MI Bayesian Optimization Algorithm for Predicting Drug–Target Interactions
Marina Yamasawa, GPU/MPI Parallelization of Metagenomic Sequence Homology Search Tool and Its Large-scale Application to Oral Microbiota Analysis Associated with Periodontal Disease
Yicong Huang, Membrane permeation prediction of cyclic peptides using enhancing sampling molecular dynamics simulation
Rikuto Kubota, Development of efficient protein-ligand docking method for virtual screening by reuse of fragments
Hiroki Watanabe, MEGADOCK-Web-Mito: a database of computer-predicted mitochondrial protein-protein interactions
Toshitaka Tanebe, Machine learning based compound activity prediction using binding pocket information