Welcome to the Cognitive Data Science (CoDS) Group at the Computer Science and Software Engineering department at Concordia University. CoDS focuses on developing systems for Linked Data Science on federated and big datasets. We aim to integrate our systems into existing data science platforms to support cognitive data science, where the scientific community can automatically discover and learn about each other’s work. The broad areas of research in the group include knowledge graphs, parallel/distributed systems, data management, and graph neural networks.
The Research Program and Objectives
The group objective is to build a world-class research program in the field of Linked Data Science. One of the key concepts to enable our vision is to abstract from syntactical differences of existing platforms and instead focus on the semantics of datasets, artifacts, and pipelines. Once we understand the semantics, we can more easily identify similar or matching artifacts and combine them in a federated manner. At CoDS, we use knowledge graph technologies to retain a maximal degree of flexibility by capturing metadata and semantics in a flexible graph format.
Our systems will help
- Discover and extract relevant data and pipelines.
- Enable reproducibility of experimental results with ease.
- Support efficient discovery of the most recent insights related to a dataset.
- Enable scientists to reuse and combine existing data science pipelines in novel ways.
- Enable scientists to collaborate more effectively regardless of the data science platforms they use.
- Encourage innovative applications to automate several aspects of data science based on the most recent data science experimentation.