On training list

The Artificial Intelligence → Detectors section displays a list of models currently being trained using client log streams. These models are automatically deployed and trained by the platform whenever the ingested logs contain the specific characteristics required by each model.

Analysts can filter the list using the following parameters:

• Detector name

• Client name

• Detectors currently generating signals

Generate signals The "Generate Signals" toggle controls whether the associated detector sends its detection signals to the Data Lake. By default, this toggle is disabled to prevent undertrained models from generating unreliable detections.

It is up to the analysts to decide when a detector should begin generating signals, based on their workload and SOC procedures.

Detections

The “Detections” control provides access to the detailed view of each model’s behavior, where analysts can review its performance, including detections relative to the volume of data processed. This section also includes monitoring and configuration options to adjust and fine-tune the model’s behavior.