Exploring accelerated machine learning for experiment data analytics
Exploring accelerated machine learning for experiment data analytics
Project goal
The project has two threads, each investigating a unique use case for the Micron Deep Learning Accelerator (a modular FPGA-based architecture). The first thread relates to the development of a real-time streaming machine inference engine prototype for the level-1 trigger of the CMS experiment.
The second thread focuses on prototyping a particle-identification system based on deep learning for the DUNE experiment. DUNE is a leading-edge, international experiment for neutrino science and proton-decay studies. It will be built in the US and is scheduled to begin operation towards the end of this decade.
Collaborators
Project background
The level-1 trigger of the CMS experiment selects relevant particle-collision events for further study, while rejecting 99.75% of collisions. This decision must be made with a fixed latency of a few microseconds. Machine-learning inference in FPGAs may be used to improve the capabilities of this system.
The DUNE experiment will consist of large arrays of sensors exposed to high-intensity neutrino beams. The use of convolutional neural networks has been shown to substantially boost particle-identification performance for such detectors. For DUNE, an FPGA solution is advantageous for processing ~ 5 TB/s of data.
Recent progress
The CMS team primarily focussed on preparing a level-1 scouting system using the Micron SB-852 FPGA processing boards to capture and process trigger-level data at 40 MHz. We developed and optimised neural networks to improve analysis performance using level-1 scouting objects. In addition, we developed a system for level-1 anomaly detection using a variational auto-encoder approach. We implemented this using the Micron deep-learning framework.
The DUNE team organised a test on the protoDUNE Single Phase detector to analyse data from cosmic rays. This was the last chance to test it before protoDUNE’s second run in 2022. The test aimed to examine the incoming triggered data using a triple AC-511 Micron FPGA as a hardware accelerator. The hardware ran an image-segmentation neural network designed to identify regions of interest. This setup was able to analyse data at a rate of ~1.42 Gb/s. The capability of the network to identify low-energy events was tested in data taken by protoDUNE with a neutron generator.
Next steps
Publications
- D. Golubovic, 40 MHz scouting with deep learning in CMS. Published on Zenodo, 2020. cern.ch/go/vJD9
- M. Popa, Deep learning for 40 MHz scouting with level-1 trigger muons for CMS at LHC run-3. Published on Zenodo, 2020. cern.ch/go/99St
Presentations
- M. J. R. Alonso, Fast inference using FPGAs food DUNE data reconstruction (7 November). Presented at 24th International Conference on Computing in High Energy and Nuclear Physics, Adelaide, 2019. cern.ch/go/bl7n
- M. J. R. Alonso, Prototyping of a DL-based Particle Identification System for the Dune Neutrino Detector (22 January). Presented at CERN openlab Technical Workshop, Geneva, 2020. cern.ch/go/zH8W
- T. O. James, FPGA-based Machine Learning Inference for CMS with the Micron Deep Learning Accelerator (22 January). Presented at CERN openlab Technical Workshop, Geneva, 2020. cern.ch/go/pM7P
- M. Rodríguez, S. A. Monsalve, P. Sala, Prototyping of a DL-based Particle Identification System for the Dune Neutrino Detector (22 January). Presented at CERN openlab Technical Workshop, Geneva, 2020. cern.ch/go/zH8W
- D. Golubovic, T. James, E. Meschi, 40 MHz Scouting with Deep Learning in CMS (22-24 April). Presented at Connecting the Dots Workshop, New Jersey, 2020. cern.ch/go/C96N