Complex materials engineering via Provenance Action Graphs: potential, structure and demonstration

Key points

• The importance comes from the requirement to synthesize technologically useful materials, especially those that are kinetically stabilized.1
• Often, it is a science that follows years of trial and error, culminating in (published) successful recipes, but is truly driven by far more (unpublished) failures.1
• The entry into the era of AI provides an opportunity to use learning algorithms to encode rules from multi-dimensional synthesis pathways, regardless of their technological usefulness, and to decode both forward prediction and inverse learning modes.1
• We will focus on several scenarios in the common synthesis process during solid-state synthesis, primarily because these reactions are kinetic in nature and require closer monitoring and control of the reaction history.1

Sources (1)

• [1]Complex materials engineering via Provenance Action Graphs: potential, structure and demonstration
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 13, 12:57 PM

  This work discusses and tests the potential of reaction graphs for material synthesis, in which events consist of reaction steps (actions) and their outputs (observations).

  The importance comes from the requirement to synthesize technologically useful materials, especially those that are kinetically stabilized.

Extractive summary: sentences quoted from the sources.