Data4AI: technical challenges to make it possible (Workshop 2.4)
SESSION DESCRIPTION & OBJECTIVES
Despite the last advances on Big Data, machine learning and computing capabilities that have enabled the latest astonishing rise of Artificial Intelligence and its presence on many industrial sectors and areas of our lives, there are still some important issues to overcome in order to eventually achieve a scenario where we could state that Data driven Artificial Intelligence is (almost) everywhere. Although some of these factors fall in social and policy domains (responsibility, transparency, ethics), there are other aspects more related to existing technical shortcomings.
In their position statement paper on “Artficial Intelligence for European Economic Competitiveness and Societal Progress”, the Big Data Value Association (BDVA) states that the success in industrial AI applications relies on the combination of five wide range technologies: Advanced data analytics, Hybrid AI, Distributed AI / Edge Analytics, Hardware optimized to AI and Multi-lingual AI.
However, the effective contribution of each of these technologies to a real Data4AI scenario is hindered by different obstacles that should be overcome for its final establishment as a reality in the industry. In this workshop, we will try to identify some of the main technical challenges that a real implementation of Data4AI faces regarding each one of the mentioned technologies, and will discuss about possible solutions to overcome those barriers.
Therefore, the workshop will be organized to foster the discussion around each one of the mentioned technologies, where relevant actors on each area will be invited to share their experiences, success stories, best practices, etc… The outcomes from the workshop will be used as contributions to the BDVA Position Statement Paper, and to extend its content about potential challenges and solutions on each of the core technologies identified for Data4AI.
- Introduction to the session
- Presentation of the BDVA Position Statement Paper on AI. Wide range technologies around Data4AI
- Panels for discussion about technical challenges around each Data4AI technologies (including representatives from PPP projects)
- Advanced data analytics
- Hybrid AI
- Distributed AI / Edge Analytics
- Hardware optimized to AI
- Multi-lingual AI
- Perspective from other initiatives – Robotics, ECSO, AIOTI, HPC… (TBD):
- Similarities and differences
- Potential collaborations
- Wrap-up / conclusions
- Sonja Zillner (Siemens)
- Ed Curry (Insight)
- Jon Ander Gómez (Universidad Politécnica de Valencia)
- Cross-CPP project (TBD)
- FourByThree project (Constantino Roldán, Tekniker)
- BigDataStack (Dimosthenis Kyriazis, University of Piraeus)