Materials Data Platform overview: metadata, vocabulary, and repository Mikiko Tanifuji Materials Data Platform Center Div. of Materials Data and Integrated System (MaDIS) National Institute for Materials Science RDA2020 RDA Breakout 6: Materials IG session: Data Infrastructure for Collaborations in Materials Research, November 12, 2020
20
Embed
Materials Data Platform overview: metadata, vocabulary ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Materials Data Platform overview: metadata, vocabulary, and repositoryMik iko Tani fu j i
Mater ia l s Da ta P la t fo rm Cen te r
D iv. o f Ma te r i a l s Da ta and In teg ra ted Sys tem (MaDIS)
Na t iona l Ins t i t u te fo r Ma te r i a l s Sc ience
RDA2020 RDA Breakout 6: Materials IG session: Data Infrastructure for Collaborations in Materials Research, November 12, 2020
2
Who we are
NIMS National Institute for Materials ScienceEstablished in 2001 by merger of two national institutes: (metals + inorganic materials) Now covers materials in general
MaDIS Research & Services Division of Materials Data and Integrated SystemEstablished in 2017 to focus on materials data and integration
DPFC Materials Data Platform CenterBudget 2017 – 2020, 3 billion yen
3
Data-driven people at MaDIS
66 staff at Materials Data Platform Center
Data Science• スパースモデリング(モデル選択)• 画像解析〜パターン認識、深層学習• 回帰技術〜機械学習、ベイズ推定• 最適化技術〜能動学習等• 自然言語処理
Data infrastructure• Data structure and modeling• Data curation• Data collection and FAIRable• Data mining from publications• Data system technology and
development
Who we are: a Facebook
4
Materials Data Platform at NIMS
Createthe data
Usethe data
Storethe data
Publishthe data
Text/Data mining
Experiments &Calculations
Analysis &Integration
Repository
Store and manage
NIMS NOW, 19 (1), 2019
5
Materials Data Platform overview
Publishing, data linking,open science
HPC Server
Analysis environment
Userauthentication
Data collection
Building advanced databases
Large-scale facilities
Analyses and materials integration
Closed data Open data
Collaboration
Industry Academia
Publishing,Federation with other repositories
Text and data mining
MatNaviDatabase
DataCollection
System IoTdata transfer
system
Otherdatabase
M-DaCData Conv
Tools
Machinelearningsystem
SIPMint
system
Research Data
Management
Common metadata
model
APIFramework
MaterialsVocabulary
Wiki
Data Management
PlanData policy
MaterialsData
Repository
Image credit: Koji101, pngimg.com (CC)
Data Provider Data Center Data Science
Academia-Private sectors
Data CloudMaterials data hub(2021 - )
6
Four actions mapped to the platform components
Createthe data
Usethe data
Storethe data
Publishthe data
Text and data mining
DataCollection
System IoTdata transfer
system
Materialsintegration
Analysis environment
MatNaviDatabase
Otherdatabase
Research Data
Management
MaterialsData
Repository
Common metadata
model MaterialsVocabulary
Wiki
Servercluster
Data policy
User auth
7
DICE Common Message Format
Characterization metadata
Method,Environment…
Specimen metadata
Material type,Structure…
Propertymetadata
Physical properties,Units…
Synthesis/Processmetadata
Processed date,Temperature…
Calculationmetadata
Computer software,Version…
Characterizationprimary params
Specimen primary params
Property primary params
Synthesis/Processprimary params
Calculationprimary params
Data DataData Data Data
Mandatory metadata
Domain-specific metadata
Primary parameters
Implementedas data model
Save as files
METADATA
DATA
Common metadata
model
Bibliographic metadata Administrative metadata Subject material
+
+
After lots of discussions (still ongoing), schema file published at https://dice.nims.go.jp/