Hi all, I wanted to let you know about the "ml-enabler", we have been starting to develop. It is a piece of software with the goal to provide the glue between machine learning input and consuming applications around OSM. The ambitious idea is to provide a kind of technical registry and facilitator for machine learning models for the OpenStreetMap ecosystem - One API that OSM gets all the ml data from.
ML models can integrate into the ml-enabler, then to be used by more than one application. Instead of programming repeatedly interfaces for each data source and each software, machine learning models can be integrated once into the ml-enabler and then be used by all applications through a consistent API. On the other side, mapping applications can be programmed connect to the ml-enabler to obtain access to data from all available ML models. Instead of integration several data sources into one consuming software, the ml-enabler can be used to connect to many models by only implementing one API. Now, a first working implementation is available on the https://github.com/hotosm/ml-enabler and https://github.com/hotosm/ml-enabler-cli repositories. This consists in a proof of concept with two machine learning models (looking glass and MS buildings), connecting to a first consuming application, the Tasking Manager. Of course it would be nice to receive your feedback and thoughts! And if you are dealing with models or want to consume any in your application, you might want to check out the ml-enabler. More information can be found in a related blog post: https://www.hotosm.org/updates/the-machine-learning-enabler/ We are going to do a technical introduction and round of conversation next week. You are welcome to join: https://www.eventbrite.com/e/hot-tech-team-updates-ml-enabler-tickets-64672382838 Thanks, Felix -- machine-learning mailing list [email protected] https://lists.openstreetmap.org/listinfo/machine-learning
