1st Real Time Analysis Workshop


Real-time analysis consists of processing and drawing inferences from data which are too large to be recorded to permanent storage. Both industry and fundamental research increasingly require reliable real-time analysis, whether the objective is to search for new particles, measure properties of known particles with unprecedented precision, control and calibrate experiments in real time, enable self-driving cars, smart cities, fraud detection, monitoring of medical interventions, etc. 

Today, however, real-time analysis is not yet seen as a scientific domain in its own right. The fact that expertise in the different facets of real-time analysis is dispersed across many domains and disciplines slows down both conceptual and practical advances in this area. Nevertheless, important leaps forward have occured in the past years, notably at the Large Hadron Collider where experimental teams have demonstrated the ability to align and calibrate their detectors in real time with optimal fidelity, and use the data so processed to make real-time inferences about particles and processes which would have only been accessible offline in the past.

The objective of this workshop is to bring together the community of high-energy physicists who are driving the developments of real-time analysis within their domain with key real-time analysis specialists from industry. By coalescing around concrete problems, and searching for common solutions, we will spark new collaborations and bring this disparate community closer together.

We will tackle four major thematic areas for real-time analysis at this workshop

1. Partial event building and persistency

2. Real-time machine learning

3. Real-time alignment and calibraition

4. Efficient use of modern CPU architectures, vectorization, and cross-architecture real-time programming.

The workshop will be held at, and is sponsored by, the Institut Pascal, a purpose built scientific hub which is part of the Université Paris Saclay. It is part of a three event series called Learning To Discover which will take place during 2019 and 2020. Many of the organizers of this event are part of the SMARTHEP consortium, and an LHCb collaboration hackathon will be hosted during the second half of the workshop in collaboration with the LHCb Computing project.