DevoLearn: a platform for open Developmental Data Science, Machine Learning, and Education

DevoLearn (http://github.com/devolearn) is an outcome of the DevoWorm group’s Machine Learning and Computational Modeling activities, sponsored in part by INCF’s involvement in the Google Summer of Code program. We will give a brief tour of the platform, in its current form as well as future development. Our starting point will take a look at the DevoLearn pre-trained deep learning model. DevoLearn is currently in version 0.3.0, and is also hosted at PyPi (https://pypi.org/project/devolearn/). The DevoWormAI (https://devoworm.github.io/ DevoWormAi/index.html) model library, DevoZoo (https://devoworm.github.io/devozoo/ index.html) curated data library, and additional resources such as data science demos and theory-building resources round out our currently-implemented capabilities. Finally, we will discuss future development, which focuses on the creation of an Epistemological Directory (ED - https://github.com/OREL-group/Epistemological-Directories) for embryogenesis and other topics in developmental biology. EDs are informatics tools that allow newcomers to engage with interdisciplinary education and specialized, field-specific concepts. EDs include key readings, literature reviews, discussion streams, and note taking systems. To encourage sustainability, the maintenance and revision of DevoLearn will be tied to activities of the DevoWorm group (http://devoworm.weebly.com), where we engage students from around the world of topics spanning theoretical biology, neural development, and methodological innovation.