GREMLIN — Finding What Machines Don't Tell You
Year: 2026
Type: article
Abstract: Passive RF diagnostics and machine-state inference at FAIR. A 10-minute show-and-tell at the European GNU Radio Days 2026, Rennes. GREMLIN explores how much can be learned about complex small- and large-scale infrastructure and technical systems from the emitted RF signals they already produce. The talk presents ongoing work at FAIR as part of the EU iRIS project, combining high-rate RF signal acquisition, modern C++/DSP, time-series replay and machine-learning-assisted inference to detect hidden states, anomalies and correlations across distributed equipment, using maintainable, software-defined instrumentation based on GNU Radio 4.0. The record holds the slides as PDF, one page per slide and one page per step, and the gr4-present deck itself (zip, with speaker notes). Live slides in the browser: https://fair-acc.github.io/gremlin/talks/viewer/a65dc67/?presentation=../../20261116_EU_GR_Days
Institution: GSI Helmholtz Centre for Heavy Ion Research
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: European Commission
OpenAlex ID: W7226027889
