An object-oriented framework to enable workflow evolution across materials acceleration platforms

What is missing in autonomous discovery: open challenges for the community

“…Scaling SDLs will require actional workow systems-mediated perhaps by powerful workow languages-designed to enable a description of the requirements for each sub-task in a workow as well as the specics needed to connect intermediate products from one sub-task to the input of subsequent sub-tasks. 28,34,44,180,181 Intelligent search or orchestration that scales is a second opportunity for soware innovation in an ongoing area of rich study. This requires a capacity for evaluating incoming results from experiments carried out in parallel on the work oor, and converting this new data into new experimental queries.…”

“…While we will rst describe the current state of the eld, this is by no means a comprehensive review, and we encourage the reader toward reviews and perspectives of SDLs and autonomous experimentation. [1][2][3][4][5][6][7][8][9][10][11][12][13][14]16,[23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38] Following this, we will turn our attention to barriers and opportunities associated with data, hardware, knowledge generation, scaling, education, and ethics. As the eld of autonomous experimentation grows and SDLs become more common, we hope to see rapid growth in scientic discovery.…”

“…Unfortunately, constructing these relationships requires some degree of standardization and is highly discipline dependent. Attempts at universal materials formats 28,[49][50][51][52] and ontologies 53,54 exist to address this challenge, with much work remaining. Crystallography provides an excellent example of standardizing data formats for interoperability (protein data bank and crystallographic information les).…”

The laboratory of Babel: highlighting community needs for integrated materials data management

“…[1][2][3] A dizzying array of recent research has contributed tools that enable this paradigm shi, including new open hardware platforms, 4 optimization and experiment planning methods, 5 and methods for sharing procedures across different laboratories. [6][7][8] This newfound ability to generate vast troves of experimental data comes as new machine learning and data science methods build off that data, 9,10 turning it into a rstclass research product in itself. 11 However, comparably little effort has been expended on systems to collect, organize, store, and share this data effectively.…”

“…We intentionally discuss these topics at a high level and ground them in examples, rather than getting into details of technical implementations. While many recent works have explored various aspects of the data management problem, 13,14 and several projects implement isolated items that are needed to make data management work, 8,15 we believe that discussion about how disparate pieces of data management tooling t together to form an integrated system is missing. Our hope is to start an accessible conversation around what our data management systems should do, not how they go about doing this.…”