Skip to content
DraconiticResearch
Search Sources API
197,787 studies31 sourcesUpdated 52 min ago
PreprintarXiv licencearXiv

Real-Time Plasma State Prediction via FPGA-Accelerated Quantized Recurrent Probabilistic Neural Networks

  • Daniel Gaytan-Villarreal
  • Aiken Xie
  • Tu Pham
  • Rohit Sonker
  • Chiara Amendola
  • Matteo Cremonesi
  • Cong Hao
  • Jeff Schneider
Open at sourcePDFDOICite

Abstract

Real time plasma state estimation for control of Tokamak devices are challenging due to the stringent latency requirements of the plasma control system (PCS). We present an end-to-end workflow for deploying a recurrent probabilistic neural network (RPNN) on FPGA hardware. We combine architecture size reduction with quantization-aware training via QKeras. The model is then synthesized using hls4ml, targeting a Xilinx Alveo U50 device. We report a design that fits comfortably within all four resource budgets (DSP, LUT, FF, BRAM) at deterministic sub-10~μs single-timestep latency, meeting the requirements for real-time inference inside a model-predictive-control-style plasma control loop.

Related

  1. Nuclear mass prediction using bidirectional recurrent neural networks with isotopic and isotonic chain correlations
    P. Li, Y. F. Niu, F. Q. Chen +1
    arXivPreprintarXiv licenceFull textNuclear physicsarXiv
  2. Density Accelerated Relaxation in Active Passive Colloidal Networks with Competing Interactions
    Pronay Dutta, Hima Nagamanasa Kandula
    arXivPreprintCC BY 4.0Full textCondensed matterarXiv
  3. Unsupervised and probabilistic learning with Contrastive Local Learning Networks: The Restricted Kirchhoff Machine
    Marcelo Guzman, Simone Ciarella, Andrea Liu
    Proceedings of the National Academy of Sciences of the United States of AmericaPeer-reviewedCC BY-NC-NDComputational physicsCondensed matterHAL
  4. Embedding stochastic dynamics in latent manifolds of recurrent neural networks
    Ramón Nartallo-Kaluarachchi, Renaud Lambiotte, Alain Goriely
    Physical Review ResearchPeer-reviewedCC BY 4.0Atomic, molecular & optical physicsCrossref
  5. Dynamic probabilistic decision networks
    V. I. Yukalov, E. P. Yukalova
    arXivPreprintCC BY 4.0Full textStatistical & nonlinear physicsarXiv

Cite

BibTeXRIS
BibTeX
@misc{gaytanvillarreal2026realtime,
  title = {{Real-Time Plasma State Prediction via FPGA-Accelerated Quantized Recurrent Probabilistic Neural Networks}},
  author = {Daniel Gaytan-Villarreal and Aiken Xie and Tu Pham and Rohit Sonker and Chiara Amendola and Matteo Cremonesi and Cong Hao and Jeff Schneider},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.23141},
  eprint = {2609.23141},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.23141},
}
RIS
TY  - UNPB
TI  - Real-Time Plasma State Prediction via FPGA-Accelerated Quantized Recurrent Probabilistic Neural Networks
AU  - Daniel Gaytan-Villarreal
AU  - Aiken Xie
AU  - Tu Pham
AU  - Rohit Sonker
AU  - Chiara Amendola
AU  - Matteo Cremonesi
AU  - Cong Hao
AU  - Jeff Schneider
PY  - 2026
DA  - 2026/09/19
PB  - arXiv
DO  - 10.48550/arxiv.2609.23141
UR  - https://arxiv.org/abs/2609.23141
AB  - Real time plasma state estimation for control of Tokamak devices are challenging due to the stringent latency requirements of the plasma control system (PCS). We present an end-to-end workflow for deploying a recurrent probabilistic neural network (RPNN) on FPGA hardware. We combine architecture size reduction with quantization-aware training via QKeras. The model is then synthesized using hls4ml, targeting a Xilinx Alveo U50 device. We report a design that fits comfortably within all four resource budgets (DSP, LUT, FF, BRAM) at deterministic sub-10~μs single-timestep latency, meeting the requirements for real-time inference inside a model-predictive-control-style plasma control loop.
ER  -