Get Started
Home
Topics
Search
Library
Research questionHow can spaceborne FPGA neural-network accelerators detect and recover from radiation-induced silent output corruption?Proton radiation can leave a Linux-managed accelerator running while its neural-network outputs become persistently incorrect. Ordinary liveness checks and sampled hardware diagnostics may therefore miss failures that affect inference content.
AI
Machine Learning
Research Paper
Technology
Latest papersRecent research connected to this question, newest first.Proton Irradiation Characterization of an Open-Source ML Accelerator on a Zynq UltraScale+ MPSoCApplies to open-source, register-transfer-level-accessible neural-network accelerators deployed on FPGA-SoCs for spaceborne inference. The reported evidence covers an unmitigated accelerator under 20–58 MeV proton irradiation and does not validate a particular mitigation or establish the cause of field-associated events.research paper · Sep 4, 2026
Related questions
How can quantum neural networks prevent simulator-era postprocessing from silently discarding hardware measurements?How can lunar robots maintain accurate, bounded-latency instance segmentation under low light, tight compute, and radiation faults?How can edge analog in-memory inference reduce energy while certifying errors from heterogeneous, imperfect accelerators?How can we estimate a strongly coupled reactor’s full internal state from three noisy external measurements across operating and accident scenarios?