The Center for Dynamic Causality is honored to announce that it has contributed two new high-level Domain-Specific Language (DSL) layers to the DeepCausality Project. The Causal Discovery Language simplifies the process of causal discovery, and the Causal Flow Language simplifies and streamlines the development of complex, multi-stage, dynamic causal processes. Together, they form a new foundation for causal discovery and execution in the DeepCausality Project.
Causal Discovery Language
The Causal Discovery Language has been rebuilt around a type-encoded pipeline that hosts multiple state-of-the-art algorithms. SURD and the new Bayesian Root-Cause Discovery (BRCD) algorithms now coexist as type-encoded peers. BRCD has been verified against the RCAEval benchmark. Details are written up in the DeepCausality announcement, The Causal Discovery Language, Rebuilt.
Causal Flow Language
The Causal Flow Language is the API that developers use to write dynamic causal processes. Sequential composition, bounded iteration, and branching become first-class constructs in dynamic causality. Closed-loop examples such as corrective lane keeping now fit in a handful of readable lines. The full write-up is available as The Causal Flow Language on the DeepCausality blog.
About
The Center for Dynamic Causality is an independent research organization dedicated to the study of dynamic causality. Insights developed at the Center, once verified, are contributed back to the DeepCausality project.
DeepCausality is a dynamic-causality framework that enables dynamic causal reasoning in Rust, hosted at the LF AI & Data Foundation. For more information about DeepCausality, please visit the project website.
The LF AI & Data Foundation supports an open artificial intelligence (AI) and data community and drives open-source innovation in the AI and data domains by enabling collaboration and the creation of new opportunities for all members of the community. For more information, please visit lfaidata.foundation.