live_detection events from the inference looplive_detection events from the A100 inference loopReal episodes arrive from the fleet, get chunked into ten-second - windows, and a deployed model labels each window in flight. The - heavy models can offload inference to an A100 - so the receiver never blocks on a forward pass — predictions - stream back as they finish.
-Each row on the stage is a host; each cell is one ten-second
- window painted by the model's predicted phase. A clean run
- cruises blue; an attack profile pushes the lane through
- armed → infecting →
- infected_running. When ground truth catches up,
- mismatched cells get a hatched overlay so you can spot where
- the model disagrees with the orchestrator. The callout below
- holds the most recent prediction with model name,
- confidence, and round-trip latency.
The A100 runs inference against incoming + ten-second windows from the fleet. Each row on the stage is + one trained model doing live prediction; each cell + is its phase verdict on a freshly-arrived window, painted + by the predicted phase.
+Read the lanes side-by-side as a model-agreement check:
+ when the recurrent family (RNN / GRU / LSTM) all flip to
+ infecting at the same time, that's strong
+ evidence the host actually is. When ground truth from
+ labels.jsonl catches up, mismatched cells get
+ a hatched overlay and the running hit-rate ticks. The
+ callout below holds the most recent prediction with model
+ name, A100 round-trip latency, and confidence.