PORRITTINC
( LAYTON, UT )
DIRECTION 01 · R&D VISION

Spiking
anomaly
detection.

Conventional anomaly detection samples a tag on a scan cycle and compares it to a threshold. A spiking network watches the change itself — reporting only when the signal does something, at a fraction of the power.

← Back to Neuromorphic Compute

Why spikes beat scans
PRINCIPLE 01

Report change, not state

Event-driven sensors emit only on change. A quiet plant costs almost nothing to watch; a plant doing something interesting gets full attention immediately.

μs
Detection latency
PRINCIPLE 02

Learn the signature

Spiking networks classify the temporal pattern of an event — the shape of a pump cavitating, not just the fact that vibration crossed a number.

mW
Node power
PRINCIPLE 03

Stay on the device

Classification happens where the sensor is. No cloud round-trip, no bandwidth bill, no dependency on a network that may not exist.

EDGE
On-device

READ THE POSITION PAPER →

μs
Detection latency
mW
Power envelope
EDGE
On-device
R&D
Vision stage

Catch it early.

This is a research direction, not a shipping product. We are stating where we are going.

CONTACT TIMOTHY → Back to Neuromorphic Compute →