Por João David (LASIGE, DI/FCUL).
Current Machine Learning (ML) pipeline orchestration tools rely on pipeline execution for the detection of errors. An error found in a pipeline could result in the waste of time and resources depending on the size of the datasets used, the operations applied on them as well as the length of the pipeline.
This talk presents the progress made on the development of MLVP, a tool that uses a Visual Programming Language (VPL) for the creation of ML pipelines while statically verifying them. The tool allows for the compilation of the pipeline into executable source code.
Bio: João David is a Masters student at the Faculty of Sciences, University of Lisbon, and a student researcher at LASIGE. He received his BSc in Computer Science from the same faculty. His current research focuses on the development of a type system to statically verify the validity of machine learning pipelines.
Transmissão em direto via Zoom.
