Synthesis Action

Synthesis action research focuses on automatically extracting and representing the procedural steps involved in various synthesis processes, primarily in materials science and robotics. Current efforts concentrate on developing standardized ontologies and large-scale annotated datasets to train machine learning models, often employing neural networks, for tasks like sentence classification, named entity recognition, and relation extraction from scientific literature or robotic interaction scenarios. This work aims to improve reproducibility, automate synthesis procedures, and ultimately accelerate scientific discovery and technological advancements in fields ranging from materials synthesis to human-robot collaboration.