Artificial Intelligence Domain Specific Language (AI-DSL) for autonomous interoperability between AI services.
The general idea of the AI-DSL is to provide
A simple and powerful language to express AI service assemblages, as well as to formalize mathematical properties, both crisp and statistical, to be met by these AI service assemblages. These properties may pertain to algorithmic behaviors, interactions with other servies as well as resource usage, computational, financial or otherwise.
A tool set for not only verifying the validity of such assemblages but also automatically create such assemblages. Basically, if a user can formalize precisely enough the desired function, then the AI-DSL tool set should be able to automatically fetch and combine the right AI services to deliver that function.
Up to now the work has been mostly exploratory resulting into prototypes and experimental code covering various aspects of the AI-DSL as opposed to a final product. This is justified by the fact that it is an ambitious project and requires a fair bit of research and development.
Progress has been taking place into phases. Details can be found in technical reports written at the end of each phase. An overview of is given below.
doc
: contain technical reports and other documentationexperimental
: contain a number of experiments, ranging from
representing and proving properties using dependent types,
performing program synthesis to type checking protobuf
specifications, and more.ontology
: experiment representing SingularityNET platform
knowledge in SUO-KIF format.snet-marketplace-space
: scripts to build an atomspace of the
SingularityNET Marketplace, as well as file dumps in MeTTa and JSON
formats.