September_AMP_Digital

HIGHL IGHTS A D V A N C E D M A T E R I A L S & P R O C E S S E S | S E P T E M B E R 2 0 2 0 5 8 ASM DATA ECOSYSTEM INITIATIVE ASM International has embarked on a journey to bet- ter serve members—moving us from renewal, to realiza- tion, and ultimately, to growth. The improvements we have delivered demonstrate the sound foundation of the stra- tegic plan. As we continue on this journey, we are mindful of the rapid changes in our world. Newer technologies like artificial intelligence (AI) and machine learning (ML) are transforming design and optimization activities—beyond traditional Integrated Computational Materials Engineering (ICME), embodying the Materials 4.0 capabilities needed to serve the Industry 4.0 world. This shift impacts our mem- bers’ ways of working across the product lifecycle. Think of our new reality in two interdependent buckets—models and data. These frame the ASM Data Ecosystem. Let’s start with models related to materials. George Box, a famous British statistician, is credited with the phrase “All models are wrong, and some are useful.” ICME dates back into the 1980s, and as computational power has increased and storage costs for data have decreased, the sophistication of models to design materials and lever- age them in design has proven to be very “useful,” and by 2011, the Materials Genome Initiative became a reality. The emergence of machine learning in design and optimization has created discontinuities the market is only beginning to exploit. Combined, ICME and ML have resulted in materials design speed and efficacy that’s nothing short of spectac- ular. That said, the industrial ecosystem’s ability to deploy these new materials is advancing, albeit, at a slower pace. We’re confident that pace will accelerate soon, because market pressures for optimization will require it. So let’s now consider materials data. Clive Humby has been credited with saying “data is the new oil.” Only part of his point is conveyed in that often-quoted sound bite. His primary point is much more thoughtful—“data is valuable, but if unrefined, it cannot really be used.” Refined materi- als data and related analytical tools enable digital threads spanning activities from design and R&D, to in-production, and in-service performance outcomes. Refined materials data enables a new way of working in which ASM members and stakeholders convert these data into realized value. It’s more predictive and less experimental. Sowhat is the ASMData Ecosystem? It’s a natural exten- sion of ASM’s core mission as the Materials Information soci- ety. The same trusted brand—with peer-reviewed content, high quality educational products and events—will take the same approach into the world of materials data and related analytics. The Data Ecosystem architecture (see diagram, left side) shows three layers and an educational capability alongside. The top layer is the Software as a Service (SaaS) Cloud, in which vetted commercial and freeware platforms for materials and engineering simulation are securely provi- sioned to subscribers. This SaaS cloud will include machine learning platform tools where these subscribers can create their own models from ASM datasets or their own. ASM Data Ecosystem Initiative Ray Fryan, ASM Executive Director – New Product Development The middle layer is the Data Lake, where high quality, curated datasets are stored and offered for use to subscrib- ers. These datasets can be used with ASM provisioned tools or the user’s own software. In addition, this lake will be seg- mented and secured in a way that the datasets are accessed only by authorized subscribers. The bottom layer—the “data factory” that feeds the lake—will be a portfolio of trusted experimentally and/or simulation-generated datasets. To wear the ASM brand, these datasets will be tested and refined via a robust, peer-reviewed curation process. Finally, our members are already seeing the impor- tance of thoughtfully managing this “new oil” of materi- als data. Accordingly, we’re initiating the development of trusted, objective data management education. Our Mate- rials Data Management School will leverage ASM’s deep net- work of materials science and data science expertise, with trusted, objective, educational products and services. Our members and stakeholders will greatly benefit from these new ASM capabilities. Members who want to start the Materials 4.0 journey will get access to the world of ML, ICME, Materials and Engineering Simulation, and Materi- als Data without a big up-front investment. All members will have the opportunity to “test drive” simulation tools before committing to longer term arrangements. Most importantly, all subscribers will benefit from a data repository built by the most trusted brand for materials information—ASM International. Our ASM International Operations Team members are already evaluating initial simulation, data, education, and infrastructure options as we build out a prototype version of this Ecosystem. As we scale-up, we invite your input to Left image: Architecture of ASM Data Ecosystem. Note, the specific tools or data sources shown are for Illustration purposes only.